Академический Документы
Профессиональный Документы
Культура Документы
Third round
1995
Problem 1. Let p and q be positive numbers such that the parabola y = x2 ; 2px + q has
no common point with the x-axis. Prove that there exist points A and B on the parabola such
that the segment AB is parallel to the x-axis and 6 AOB = 90 (O is the coordinate origin) if
and only if p2 < q 14 . Find the values of p and q for which the points A and B are de
ned in
an unique way.
Solution. Since the parabola has no common point with the x-axis, then the roots of the
equation x2 ; 2px + q = 0 are not real and hence p2 < q . Let the points A(x1 y0 ) and B (x2 y0)
(Figure 1) be with the required properties. Then x1 and x2 are the roots of the equation
x2 ; 2px + q ; y0 = 0 and y0 > q ; p2, because the vertex of the parabola has coordinates
(p q ; p2 ). On the other hand OA2 = x21 + y02 , OB 2 = x22 + y02 , AB 2 = (x1 ; x2 )2 and it follows
from the Pythagorean theorem that y02 + x1x2 = 0. But x1 x2 = q ; y0 and thus y02 ; y0 + q = 0.
Consequently the existence of the points A and B is equivalent to the assertion that the equation
f (y) = y2 ; y + q = 0 has a solution y0 > q ; p2. (A and B are de
ned in an unique way if
this is the only solution.) A necessary condition is that the discriminant of the equation is not
negative, i.e. q 41 . The last condition is sucient because f (q ; p2) = (q ; p2 ) + p2 > 0 and
1 > 1 q ; p2 . The corresponding solution y is unique i q = 1 .
2 4 0
4
Figure 1. Figure 2.
1
Problem 2. Let A1A2A3A4A5A6A7, B1B2B3 B4B5B6B7, C1C2C3C4C5C6C7 be regular
heptagons with areas SA , SB and SC , respectively. Let A1 A2 = B1 B3 = C1C4. Prove that
1 < SB + SC < 2 ; p2:
2 SA
Solution. Let A1 A2 = a, A1 A3 = b, A1 A4 = c (Figure 2). By the Ptolomeus theorem for the
quadrangle A1 A3 A4 A5 it follows that ab + ac = bc, i.e. a + a = 1. Since 4A1A2 A3
= 4B1 B2 B3 ,
b c
then B1 B2 = a and hence B B = a . Analogously C C = a . Therefore SB + SC = a + a .
2 2 2 2
B12B3 2 b 1 2
b 1 2
c SA b2 c2
a a 1 a a 1 a a
Then + > ( + )2 = (equality is not possible because 6= ). On the other hand
b c 2 b c 2 b c
a2 + a2 = a + a 2 ; 2a2 = 1 ; 2a2 :
b2 c2 b c bc bc (1)
a 2 sin2 7 1 2 <
By the sine theorem we get bc = = . Since cos
sin2 7 sin 47 4 cos 27 (1 + cos 27 ) 7
p p
cos 4 = 22 , then abc > p 1 p = 2 ; 1. From here and from (1) we get the right hand
2
2 2
4 2 (1 + 2 )
side inequality of the problem.
Problem 3. Let n > 1 be an integer. Find the number of the permutations (a1 a2 : : : an) of
the numbers 1 2 : : : n with the following property: there exists only one index i 2 f1 2 : : : n ;
1g such that ai > ai+1 .
Solution. Denote by pn the number of the permutations with the given properties. Obviously,
p1 = 0 and p2 = 1. Let n 2. The number of the permutations with an = n is equal to pn;1.
Consider all the permutations
! (a1 a2 : : : an ) with ai = n, where 1 i n ; 1 is
xed. Their
number is ni ;; 11 . Consequently
nX !
;1
n;1 n;1
pn = pn;1 +
i=1 i ; 1 = pn;1 + 2 ; 1:
From here
pn = (2n;1 ; 1) + (2n;2 ; 1) + + (2 ; 1)
= 2n ; n ; 1:
Problem 4. Let n 2 and 0 xi 1 for i = 1 2 : : : n. Prove the inequality
n
(x + x + + x ) ; (x x + x x + + x x + x x )
1 2 n 1 2 2 3 n;1 n n 1 : 2
When is there an equality?
Solution. Denote by S (x1 x2 : : : xn ) the left hand side of the inequality. This function is
linear with respect to each of the variables xi . Particularly,
S (x1 x2 : : : xn) max (S (0 x2 : : : xn) S (1 x2 : : : xn)) :
2
From here it follows by induction that it is enough to prove the inequality when all xi are
equal to 0 or 1. On the other hand for arbitrary xi we have
2S (x1 x2 : : : xn) = n ; (1 ; x1 )(1 ; x2) ; (1 ; x2)(1 ; x3 ) ; ; ( )
;(1 ; xn)(1 ; x1) ; x1x2 ; x2x3 ; ; xn x1
i.e. S (x1 x2 : : : xn ) n , when xi 2 0 1]. In the case when xi are equal to 0 or 1, the left
2 n
hand side of the last inequality is an integer. Consequently S (x1 x2 : : : xn) 2 . It follows
from ( ) that
when n is even, the equality is satis
ed i (x1 x2 : : : xn) = (0 1 0 1 : : : 0 1)
when n is odd, the equality is satis
ed i (x1 x2 : : : xn) = (x 0 1 0 1 : : : 0 1), where
x 2 0 1] is arbitrary.
Problem 5. The points A1, B1, C1 lie on the sides BC , CA, AB of the triangle ABC
respectively and the lines AA1 , BB1 , CC1 have a common point M . Prove that if the point M
is center of gravity of 4A1B1 C1, then M is the center of gravity of 4ABC .
Solution. Let M be the center of gravity of 4A1B1 C1. Let A2 be a point on MA! such
that B1 A1 C1A2 is a parallelogram. The points B2 and C2 are constructed analogously. Since
A1C1kA1B1kC1B2, then the points A2, C1, B2 are colinear and C1 is the midpoint of A2B2.
The same is true for the points A2 , B1 , C2 and C2 , A1 , B2 . We shall prove that A2 = A, B2 = B
and C2 = C , which will solve the problem.
Assume that A2 6= A and let A be between A2 and M . Then C2 is between C and M , B is
between B2 and M and consequently A2 is between A and M , which is a contradiction.
Problem 6. Find all pairs of positive integers (x y) for which xx ;+ yy is an integer which
2 2
is a divisor of 1995.
Solution. It is enough to
nd all pairs (x y ) for which x > y and x2 + y 2 = k(x ; y ), where
k divides 1995 = 3:5:7:19. We shall use the following well-known fact: if p is a prime number
of the type 4q + 3 and if it divides x2 + y 2 , then p divides x and y . (For p = 3 7 19 this can
be proved directly.) If k is divisible by 3 then x and y are divisible by 3 too. Simplifying by
9 we get an equality x21 + x21 = k1(x1 ; y1 ), where k1 divides 5:7:19. Considering 7 and 19 in
an analogous way we get an equality a2 + b2 = 5(a ; b) (it is not possible to get an equality
a2 + b2 = (a ; b)), where a > b. From here (2a ; 5)2 + (2b ; 5)2 = 50, i.e. a = 3, b = 1 or a = 2,
b = 1.
The above considerations imply that the pairs we are looking for are of the type (3c c),
(2c c), (c 3c), (c 2c), where c = 1 3 7 19 3:7 3:19 7:19 3:7:19.
3
BULGARIAN NATIONAL OLYMPIAD IN MATHEMATICS
Fourth round
1995
Problem 1. Find the number of all integers n > 1, for which the number a25 ; a is divisible
by n for every integer a.
Solution. Let n be with the required property. Then p2 (p prime) does not divide n since
p2 does not divide p25 ; p. Hence n is a product of pairwise dierent prime numbers. On the
other hand 225 ; 2 = 2:32:5:7:13:17:241. But n is not divisible by 17 and 241 because 325 ;3
(mod 17) and 325 32 (mod 241). The Fermat theorem implies that a25 a (mod p)
when p = 2 3 5 7 13. Thus n should be equal to the divisor of 2 3 5 7 13, dierent from 1.
Therefore the number we are looking for is 25 ; 1 = 31.
Problem 2. A triangle ABC with semiperimeter p is given. Points E and F lie on the
line AB and CE = CF = p. Prove that the excircle k1 of 4ABC to the side AB touches the
circumcircle k of 4EFC .
Solution. Let P and Q be the tangent points of k1 with the lines CA and CB, respectively.
Since CP = CQ = p, then the points E , P , Q and F lie on the circle with center C and radius p.
We denote by i the inversion dened by this circle. Since i(P ) = P , i(Q) = Q, then i(k1) = k1.
On the other hand i(E ) = E and i(F ) = F . Hence i(k) is the line AB . But k1 touches AB and
thus k touches k1.
Problem 3. Two players A and B take stones one after the other from a heap with n 2
stones. A begins the game and takes at least 1 stone but no more then n ; 1 stones. Each
player on his turn must take at least 1 stone but no more than the other player has taken before
him. The player who takes the last stone is the winner. Find who of the players has a winning
strategy.
Solution. Consider the pair (m l), where m is the number of the stones in the heap and l
is the maximal number of stones that could be taken by the player on turn. We must nd for
which n the position (n n ; 1) is winning (i.e. A wins) and for which n it is losing (B wins).
We shall apply the following assertion several times: If (m l) is a losing position and l1 < l,
then (m l1) is losing too.
Now we shall prove that (n n ; 1) is a losing position i n is a power of 2.
Suciency : Let n = 2k , k 1. If k = 1 then B wins on his rst move. Assume that
(2k 2k ; 1) is a losing position and let consider the position (2k+1 2k+1 ; 1). If A takes at least
2k stones on his rst move, then B wins at ones. Let A take l stones, where 1 l < 2k . By
the inductive assumption B could play in such a way that he could win the game (2k l) since
l 2k ; 1 the last move will be the move of B. After this move we get the position (2k m) with
m l, which is losing for A, according to the inductive assumption.
Necessity: It is enough to prove that if n is not a power of 2, then (n n ; 1) is a winning
position. Let n = 2 + r, where 1 r 2k ; 1. On his rst move A takes r stones and B is
k
faced to the position (2k r), which is losing for B.
1
Problem 4. The points C1, A1 and B1 lie on the sides AB, BC and CA of the equilateral
triangle ABC respectively in such a way that the inradii of the triangles C1AB1 , B1 CA1, A1 BC1
and A1 B1 C1 are equal. Prove that A1 , B1 and C1 are the midpoints of the corresponding sides.
Solution. We shall prove that BA1 = CB1 = AC1 (Figure Figure 1.
1). Assume the contrary and let BA1 CB1 > AC1 . Let
be the rotation at 1200 which center coincides with the in center
of the incircle of 4ABC . This rotation transforms the incircles
of the triangles C1BA1 , A1CB1 and B1 AC1 to the incircles of
the triangles A1CB1 , B1 AC1 and C1BA1 , respectively. Let A2 =
(A1), B2 = (B1) and C2 = (C1). It follows that BB2 < BC1
and BC2 < BA1 . But the incircles of the triangles BC1 A1 and
BC2B2 have equal radii (because (4AC1B1 ) = 4BC2B2 ), which
is a contradiction.
Let r be the radius of the incircles of the triangles C1 AB1 ,
B1CA1, A1 BC1pand A1 B1C1. From the triangle B1 AC1 we have
r = 1 ; B2 1 C1 : 33 , and from 4A1B1C1 which is equilateral we
p
have r = B1 C1 : 63 . From here B1 C1 = 12 and consequently A1, B1 , C1 are midpoints of the
corresponding sides.
Problem 5. Let A = f1 2 : : : m + ng, where m and n are positive integers and let the
function f : A ! A be dened by the equations:
f (i) = i + 1 for i = 1 2 : : : m ; 1 m + 1 : : : m + n ; 1
f (m) = 1 and f (m + n) = m + 1:
a) Prove that if m and n are odd then there exists a function g : A ! A such that g (g (a)) =
f (a) for all a 2 A.
b) Prove that if m is even then m = n i there exists a function g : A ! A such that
g (g (a)) = f (a) for all a 2 A.
Solution. a) Let m = 2p +1, n = 2q +1 and g(i) = p + i +1 for i = 1 2 : : : p g(i) = q + i +1
for i = m + 1 m + 2 : : : m + q g (2p + 1) = p + 1 g (p + 1) = 1 g (m + 2q + 1) = m + q + 1
g (m + q + 1) = m + 1. It is easy to check that g (g (a)) = f (a) for all a 2 A.
b) Let m = n and g (i) = m + i for i = 1 2 : : : m g (m + i) = i + 1 for i = 1 2 : : : m ; 1
g(2m) = 1.
For the converse let M = f1 2 : : : mg. It follows by the denition of f that the elements
of M remain in M after applying the powers of f with respect to superposition. Moreover,
these powers scoop out the whole M . The same is true for the set A n M . The function f is
bijective in TA and if there exists g verifying the condition, then g is bijective too. We shall prove
that g (M ) M = . It follows from the contrary that there exists i 2 M such that g (i) 2 M .
Consider the sequence i g (i) g 2(i) : : : and the subsequence i f (i) f 2(i) : : :. It is easy to see that
g(M ) = M . We deduce that there exists a permutation a1 a2 : : : am of elements of M , such
that g (ai) = ai+1 for i = 1 2 : : : m ; 1 g (am ) = a1 and f (a2i;1) = a2i+1 for i = 1 2 : : : s ; 1
f (a2s;1 ) = a1 , whereTm = 2s. The last contradicts to the properties of f which were mentioned
already. Thus g (M ) M = . Analogously g (A n M ) = A n M , if g (i) 2 A n M for i 2 A n M .
At last let us observe that when starting from an element of M and applying g we go to A n M ,
but when applying g for a second time we go back to M . The same is true for the set A n M .
2
From here and from the bijectivity of g it follows that M and A n M have one and the same
number of elements, i.e. n = m.
Problem 6. Let x and y be dierent real numbers such that xx ;; yy is an integer for some
n n
n ; yn
four consecutive positive integers n. Prove that xx ; y is integer for all positive integers n.
Solution. Let tn = xx ;; yy . Then tn+2 + b:tn+1 + c:tn = 0 for b = ;(x + y), c = xy, where
n n
3
BULGARIAN NATIONAL OLYMPIAD IN MATHEMATICS
Third round
1996
Problem 1. Prove that for all positive integers n 3 there exist an odd positive integers
x and y , such that
n n
7x2 + y 2 = 2 :
n n
n
Solution. If n = 3 we have x3 = y3 = 1.
Suppose that for an integer n 3 there are odd positive integers x , y , such that 7x2 + y 2 =
n n n n
2 . We shall prove that for each pair
n
X = x +2 y Y = j7x 2; y j and X = jx ;2 y j Y = 7x 2+ y
n n n n n n n n
we have 7X 2 + Y 2 = 2 +1 . Indeed,
n
2 2
7 x 2 y + 7x 2 y = 2 7x2 + y 2 = 2 2 = 2 +1 :
n n n n
n n
n n
Since x and y are odd, i.e. x = 2k + 1 and y = 2l + 1 (k, l are integers), then
x + y = k + l + 1 and jx ; y j = jk ; lj, which shows that one of the numbers x + y
n n n n
n n n n n n
2 2 2
and j x
n ; y jn
is odd. Thus, for n + 1 there are odd natural numbers x +1 and y +1 with the
2 n n
required property.
Problem 2. The circles k1 and k2 with centers O1 and O2 respectively are externally tangent
at the point C , while the circle k with center O is externally tangent to k1 and k2 . Let ` be
the common tangent of k1 and k2 at the point C and let AB be the diameter of k, which is
perpendicular to `, the points A and O1 lie in one and the same semiplane with respect to the
line `. Prove that the lines AO2, BO1 and ` have a common point.
Figure 1. Figure 2.
1
Solution. Denote by r, r1 and r2 the radii of k, k1 and k2, by M and N the tangent points
of k with k1 and k2, respectively and by P the common point of ` and AB (Figure 1).
It follows from O1 O2 ? ` and AB ? ` that 4BON 4CO2N . Then CNO2 = ONB , 6 6
Figure 3.
12 x ; 12 x + 21 , we nd that the function
y = 4x3 ; 3x
has a local maximums when x = 12 . Then its maximal
value in the interval ;1 1]is the biggest
among the
1 1
r r r r
value of y = 4x ; 3x is 1.
The equality M = 1 is reached only when a = 0, b = ;3 and c = 0.
Problem 4. The real numbers a1 a2 : : : a (n 3) form an arithmetic progression. There
n
2
Find the numbers a1 a2 : : : a , if they are pairwise dierent and the biggest among them is
n
equal to 1996.
Solution. Let a1 < a2 < < a = 1996 and q be the quotient of the geometric progression
n
progression which quotient is q1 . Thus, we can assume that jq j > 1, i.e. q > 1 or q < ;1. Then
ja 1 j < ja 2 j < < ja n j, from where a 6= 0 for all i.
i i i i
More exactly, either all numbers are positive (q > 1) and then a 1 < a 2 < < a n , which
i i i
together with a1 < a2 < < a shows that a k = a , i.e. the numbers a1 a2 : : : a form
n i k n
signs alternatively (q < ;1) and then the positive ones form an increasing geometric progression
with quotient q 2 , and the order is the same as in the arithmetic progression. (The numbers
a1 a2 : : : a could not be all negative, because a = 1996 > 0.)
n n
Assume now that 3 among the numbers a1 a2 : : : a are positive. Then 0 < a 2 < a 1 <
n n; n;
Thus at most two among the numbers are positive. Analogously, at most two among the
numbers are negative. Consequently, n 4.
Let n = 4. Then a1 < a2 < 0 < a3 < a4 and 2a2 = a1 + a3 , 2a3 = a2 + a4 . But q < ;1 and
the geometric progression is either a3 a2 a4 a1 or a2 a3 a1 a4. Let it be a3 a2 a4 a1. Then
a2 = a3q, a4 = a3q2 and a1 = a3 q 3 . Thus, 2a3q = a3q3 + a3 and 2a3 = a3q + a3q2 . From here
q = 1, which contradicts to q < ;1.
So n = 3. There are two possibilities:
I. a1 < a2 < 0 < a3 = 1996. Then the geometric progression is a2, a3 = a2q, a1 = a2q2. It
follows from 2a2 = a1 + a3 that 2a2 = a2q 2 + a2 q , i.e. q 2 + q ; 2 = 0. Thus, q = ;2, ;2a2 = 1996,
a2 = ;998 and the numbers are (;3992 ;998 1996).
II. a1 < 0 < a2 < a3 = 1996. Now the geometric progression is a2, a1 = a2q, a3 = a2q2.
From 2a2 = a1 + a3 we obtain 2a2 = a2 q + a2 q 2, i.e. again q = ;2. Therefore, a3 = 4a2 = 1996
and a2 = 499. The numbers are (;998 499 1996).
Problem 5. A convex quadrilateral ABCD, for which ABC + BCD < 180 , is given. 6 6
The common point of the lines AB and CD is E . Prove that ABC = ADC if and only if 6 6
AC 2 = CD CE ; AB AE:
Solution. Let ABC = , ADC = , BAC = ' and CAD =
(Figure 4). The point A
6 6 6 6
is between E and B and the point D is between E and C . Also, AEC = + ' +
; 180 . 6
4
BULGARIAN NATIONAL OLYMPIAD IN MATHEMATICS
Fourth round
1996
Let now p and q be such prime numbers that (5 ; 2 )(5 ; 2 ) is an integer. If pj (5p ; 2p ),
p p q q
pq
then p = 3. Since 53 ; 23 = 3 3 13, then either q j (5q ; 2q ), i.e. q = 3, or q = 13. Therefore the
pairs (3,3), (3,13), (13,3) satisfy the problem condition. It remains the case when p 6= 3 q 6= 3.
Now pj (5q ; 2q ) and q j (5p ; 2p ). We can assume that p > q . It is clear that (p q ; 1) = 1 and
consequently, there are positive integers a and b, for which ap ; b(q ; 1) = 1 (Bezou theorem).
Since (q 5) = (q 2) = 1, it follows by the Fermat theorem that 5q;1 2q;1 (mod q ). From
5p 2p( (mod q ) we deduce that 5ap 2ap (mod q ) and therefore, 5b(q;1)+1 2b(q;1)+1
(mod q ). But 5b(q;1)+1 5 (mod q ) and 2b(q;1)+1 2 (mod q ). Thus, q = 3, which is a
contradiction. Finally, (p q ) = (3 3) (3 13) (13 3).
Problem 2. Find the side length of the smallest equilateral triangle in which three disks
with radii 2, 3 and 4 without common inner points can be placed.
Solution. Let in a equilateral 4ABC two disks with radii 3 and 4 without common inner
points be placed. It is clear that a line ` exists, which separates them, i.e. the disks are in
dierent semiplanes with respect to ` (Figure 1).
Figure 1. Figure 2.
This line divides the triangle into a triangle and a quadrilateral or into two triangles. In
both cases the disks can be replaced in the
gure in a way that each of them is tangent to two
of the sides of 4ABC . It is clear that the new disks have no common inner point. Let the disks
be inscribed in 6 A and 6 B of 4ABC , respectively. We translate the side BC parallelly to itself
1
towards the point A, till the disk which is inscribed in 6 B touches the disk which is inscribed in
6 A (Figure 2). Thus, we get an equilateral 4A1B1 C1 with a smaller sides, in which two disks
with radii 3 and 4 and without common inner points are placed.
Figure 3. Let A1 B1 = x, I be the incenter of 4A1B1 C1, while O1
and
x O2 be the centers of the two disks. Then A1I = B1 I =
p A1O1 = 6 B1O2 = 8. Since the disk with radius 4 is
3
inside thep4A1B1 C1, then O2 2 IB1 . Thus, B1 O2 B1 I ,
i.e. x 8 3. On the other hand O1I = px ; 6 O2I = px ;
3 3
8 O1O2 = 7 and 2 by the cosine
2 theorem for 4
x O 1O 2 I we
nd
x x x
that p ; 6 + p ; 8 + p ; 6 p ; 8 = 49.
3p 3 p3 3
Butpx 8 3, and from here x = 11 3. Consequently, AB
11 3. On thep other hand in the equilateral 4ABC with
side length 11 3 three disks with radii 2, 3 and 4 (without
common inner points) can be placed inscribing circles with
these radii in the angles of the triangle (Figure 3).
Note that the disks wit radii 3 and 4 are tangentpto each other. It follows from the above
considerations that the solution of the problem is 11 3.
Problem 3. The quadratic functions f (x) and g(x) are with real coecients and have the
following property: if the number g (x) is integer for a positive x, then the number f (x) is integer
too. Prove that there are such integers m and n, that f (x) = mg (x) + n for all real x.
q
Solution. Let g (x) = px2 + qx + r. We can assume that p > 0. Since g (x) = p(x + )2 + r ;
q2
2p 4p
q
after the variable change of x by x + 2p we reduce the problem for the following quadratic
functions
s f (x) = ax2 + bxs+ c and!g(x) = px2 + s p > 0. Letsk be such
! an integer thatsk > s and
k ; s > q . Since g k ; s = k is integer, then f k ; s = a(k ; s) + b k ; s + c
p 2p p p p p
is an integer too. Consequently, the number
s ! s !
f k + 1 ; s ; f k ; s = pb : p 1 p +a (1)
p p p k+1;s+ k;s p
is an integer for all k which are suciently big. It follows from here that ap is an integer. Indeed,
suppose that a is not an integer. If b > 0, we chose k in a way that
p
a
pp k + 1 ; s + k ; 1 < p + 1 ; ap
b :p 1 p
2
for all k, which are suciently
s big.! This is possible only when b = 0.
Let = m. Then, f k ; s = m(k ; s) + c is an integer (when k is suciently big), i.e.
a
p p
c ; ms is an integer. Let n = c ; ms. Now it is clear that f (x) = mg(x) + n for all x.
Problem 4. The sequence fan g1n=1 is de
ned by
a1 = 1 an+1 = ann + an n 1:
n
Prove that ba2n c = n when n 4 (it is denoted by bxc the integer part of the number x).
x n
Solution. Let f (x) = + . Since f (a) ; f (b) =
(a ; b)(ab ; n2 ) , it follows that the
n x abn
function f (x) is decreasing in the interval (0 n).p
Firstly, by induction we shall prove that n an p n when n 3. We have
p p n ; 1n
a1 = 1 a2 = 2 and a3 = 2, i.e. 3 a3 p2 . Let n an p
3
for an integer n 3.
n ; 1
p p
Then, an+1 = f (an ) f ( n) = np+n1 and an+1 = f (an ) f p n = p n > n+1
n;1 n;1
and thus the induction
p
nishes. p
Since an n, it remains to prove, that an < n + 1. We have an+1 = f (an )
f pn = p n when n 3. Consequently, an pn ; 1 when n 4. Then,
n;1 n;1 n;2
n ; 1 (n ; 1)2 + n2 (n ; 2) p
an+1 = f (an ) < f p = p < n+2
n;2 (n ; 1)n n ; 2
pn an < p4.n +(The
when last inequality is equivalent to 2n2(n ; 3) + 4n ; 1 > 0.) Therefore,
1 when n 4, i.e. ba2n c = n.
n
Problem 5. The quadrilateral ABCD is inscribed in a circle. The lines AB and CD meet
each other in the point E , while the diagonals AC and BD | in the point F . The circumcircles
of the triangles AFD and BFC have a second common point, which is denoted by H . Prove
that 6 EHF = 90.
Solution. Let O be the circumcenter of ABCD. We shall prove that O is the second common
point of the circumcircles of 4AHB and 4CHD. (Since AB and CD are not parallel, then
O 6= H .) After that we shall prove that the points E , H and O are colinear and 6 OHF = 90.
We shall consider the possible positions of H .
Let G be the common point of AD and BC (these lines are not parallel because the cir-
cumcircles of 4AFD and 4BFC are not tangent). It is clear that H is in the interior of
6 AGB .
1) H is in 4CGD. Then, 6 CHD = 6 CHF + 6 DHF = 180 ; 6 CBF + 180 ; 6 DAF =
360 ; CD d > 180 (from CD d < AB d and AB d + CDd < 360), which is impossible.
2) H is in 4CFD. Then, 6 CHD = 360 ; 6 CHF ; 6 DHF = 6 CBF + 6 DAF = CD d=
6 COD and 6 AHB = 6 AHF + 6 BHF = 6 ADF + 6 BCF = AB d = 6 AOB .
3) H is in 4ABF . Analogously, 6 CHD = 6 COD and 6 AHB = 6 AOB .
4) H is on AB . Again 6 CHD = 6 COD and 180 = 6 AHB = 6 AOB . Consequently, O is
the midpoint of AB .
5) H is not in 4ABG. Then 6 CHD = 6 COD and 6 AHB = 360 ; 6 AOB .
3
Note that in the cases 2), 3) and 5) the points O and H are in one and the same semiplane
with respect to the line AB and with respect to the line CD. Consequently, the points A B H
and O are concyclic. We have the same for the points C D H and O.
Now we shall prove that the line EH passes through the point O. Let this line meet the
circumcircles of 4AHB and 4CHD at the points O1 and O2 , respectively. Since these points
lie on the ray EH ! , it follows from the equalities EH:EO1 = EA:EB = EC:ED = EH:EO2
that O1 O2 O.
The points O and H are in one and the same semiplane with respect to CD and we can
assume that the quadrilateral COHD is convex. Let H be in 4CFD. Then, 6 OHF = 6 FHC ;
6 OHC = 180 ; 6 FBC ; 6 ODC = 180 ; 21 :6 COD ; (90 ; 12 :6 COD) = 90. The other
possibility for H is to be inside 6 AGB and outside the non-convex quadrilateral AFBG.
Hence 6 OHF = 6 OHC + 6 FHC = 6 ODC + 6 FBC = 90 .
Problem 6. A square table of size 7 7 with the four corner squares deleted is given.
Figure 4. a) What is the smallest number of squares which need to be colored black
so that a 5-square entirely uncolored Greek cross (Figure 4) cannot be found
on the table?
b) Prove that it is possible to writte integers in each square in a way that
the sum of the integers in each Greek cross is negative while the sum of all
integers in the square table is positive.
Solution. Denote the square in row i and column j by (i j ). Note that
a cross is uniquely determined by its central cell. The cross with central cell
(i j ) is denoted by Cij , 2 i j 6. The number of all crosses is 25.
a) The squares (1 i), (i 1), (7 i), (i 7), i = 2 : : : 6, are included in exactly one cross the
squares (2,2), (2,6), (6,2), (6,6) are included in exactly 3 crosses the squares (2 i), (i 2), (6 i),
(i 6), i = 3 4 5 | in exactly 4 crosses and
nally, the squares (i j ) 3 i j 5 are included
in exactly 5 crosses. Since C22 contains a colored square, then at least one of (1,2), (2,1), (2,2),
(2,3), (3,2) should be colored. Similarly, at least one of (2,7), (1,6), (2,6), (2,5), (3,6), at least
one of (7,2), (6,1), (6,2), (5,2), (6,3) and at least one of (6,7), (7,6), (6,6), (6,5), (5,6) are colored.
Any of them is contained in at most four crosses (the
rst two in each quintuple in 1 cross, the
third one in 3 crosses and the remaining two | in 4 crosses). Denote by x the number of the
colored squares among (i j ) 3 i j 5. The number of crosses containing a colored square is
not greater than 4:4 + 5x, whence
16 + 5x 25: (1)
Thus, x 2 and the number of the colored cells is at least 6.
Suppose that the number of the colored squares is 6. Then x = 2. Moreover, acording (1)
there exists at most one Greek cross containing more than one colored square. If two squares
with a common side or a common vertex are colored, then it is easy to check that there are two
crosses with at least two colored squares each. Therefore, the squares (4,4), (3,3), (3,5), (5,3),
(5,5) should be uncolored. With no loss of generality, let the two colored squares be (4,3) and
(4,5). But now the crosses C34 and C54 do not contain colored squares.
To prove that the minimal number of colored squares is 7, color for example the squares
(2,5), (3,2), (3,3), (4,6), (5,4), (6,2), (6,5).
b) Let us write ;5 in each of the colored squares from a) and 1 in the remaining squares.
Since every Greek cross contains a colored square, the sum of the numbers in its squares does not
exceed 1+1+1+1;5 = ;1 < 0. The sum of all numbers in the table is 7:(;5)+(45;7):1 = 3 > 0
and we are done.
4
XLV National Mathematics
Olympiad: 3rd round, April
1997
Problem 1. Find all natural numbers a, b and c such that the
roots of the equation
x2 ; 2ax + b = 0
x2 ; 2bx + c = 0
x2 ; 2cx + a = 0
are natural numbers.
Solution: Let fx1 x2g, fx3 x4g and fx5 x6g be the roots of the
rst, second and third equation respectively, and let all of them be
natural numbers.
Assume that x 2 for all i = 1 2 : : : 6. Then 2a = x1 + x2
i
of generality suppose x1 = 1, so 1 ; 2a + b = 0.
If x 2 for i = 3 4 5 6, then
i
2
Solution: Let 6 AEB = , EC = c, ED = d, ;!i = 1c ;! EC and
j = d1 ;!
;! ED. Since ABCD is an inscribed quadrilateral, CDAB =
AE = BE = k. Therefore ;! EA = kc;!
j and ;!
EB = kd ;!i . Since
CE DE
F 2 AC and F 2 BD there exist x and y such that
;!
EF = x;!
EA + (1 ; x);!
EC = xkc;!j + (1 ; x)c;!
i
and ;!
EF = y;!EB + (1 ; y);!
ED = ykd;! i + (1 ; y)d;!
j:
Comparing the coecients of ;!
i and ;!
j in these equalities gives
xkc = (1 ; y)d and ykd = (1 ; x)c. This implies x = (kkd2 ;;1)c c .
Therefore
;!
EF = k2 k; 1 (kd ; c);!
j + (kc ; d);!
i
and thus
!2
k
EF = k2 ; 1 (kd ; c)2 + (kc ; d)2 + 2(kd ; c)(kc ; d) cos :
2
MN = 1 AB ; CD
EF 2 CD AB
Problem 3. Prove that the equation
x2 + y2 + z2 + 3(x + y + z) + 5 = 0
has no solution in rational numbers.
Now 2 ; + 14 = 0, so = 21 .
n
for all n. Thus f (x0) = f (x1) = which means that f (x0) = f 21
1
for all x0 2 0 2 .
(II) x0 > 12 . Consider the following sequence:
(2) x0 x1 : : : x : : :
n
s
dened by x +1 = x ; 14 .
n n
1
f (x +1 ) = f x +1 + 4 = f (x )
n
2
n n
for all n we get that f (x0) = f 21 .
Therefore f (x) is a constant function in the interval 0 +1), and
since it is even, it is a constant for all x.
Conversely, any constant function satises the conditions of the
problem.
such that there are no two consecutive numbers in one and the same
set.
f1 = 1, f2 = 2, f3 = 4.
Divide the subsets of S having no two consecutive numbers into
n
group is f ;2 + 1. Therefore
n
f = f ;1 + f ;2 + 1:
n n n
u1 = 2 u2 = 3 u = u ;1 + u ;2 :
n n n
n
5
8
XLV National Mathematics
Olympiad: 4th round, May
1997
Problem 1. Consider
!
the polynomial
! ! !
Pn (x) = 2 + 5 x + 8 x + + 3k n+ 2 xk
n n n 2
n 2
where n 2 is a natural number and k = 3 .
;
(a) Prove that Pn+3 (x) = 3Pn+2 (x) 3Pn+1 (x) + (x + 1)Pn (x)
;
(b) Find
;
all integer numbers a such that Pn (a3) is divisible by
n
3 2 ] for all n 2.
1
Solution:
(a) Compare the coe
cients in front of xm, 0 m
n + 1 . It su
ces to show that
3 ! ! ! ! !
n+3 =3 n+2 n + 1 n n
3m + 2 3m + 2 3 3m + 2 + 3m + 2 + 3m 1 :
;
;
1
! ! !
Using the identity a +b 1 = ab + b a 1 we get that
;
! !! ! !!
n+3 n+2 2 n + 2 n + 1 +
3m + 2 ! 3m + 2 !! 3m + 2! 3m + 2
; ; ;
+ 3nm++12 ;
3 m
n
+ 2 ;
3m
n
1 =
! ! ! !
;
n + 2 n + 1
= 3m + 1 2 3m + 1 + 3m + 1 n n =
;
! !!
;
3 m 1 ;
= n + 2 n + 1 +
3m + 1 ! 3m + 1 !!
;
!
n+1 n n
;
3!m + 1 ! 3m + 1 ! 3m 1 =
; ;
;
= n3+m1 ;
n
3m ;
n
3m 1 =
;
= 0:
(b) Suppose a satises the condition of (b). Then P5(a3) =
10 + a3 is divisible by 9 and so a ; 1 (mod 3). On the contrary,
let a ; 1 (mod 3). Now a3 + 1 0 (mod 9). Since P2(a3) =
ity
sin 6 CAM + sin 6 CBM p
2
3
2
(a) if the circumscribed circle of AMC is tangent to the line AB
4
and 6 CB1M 90. Now sin 6 CAM + sin 6 CBM sin 6 CA1M +
sin 6 CB1M 2 .
p
3
3
It remains to consider the case 6 CA1M > 90, 6 CB1M 90
(the above angles could not both be obtuse). It follows from CA1M
4
(a1 b1 c1 are sides of A1B1C ). We know from (a) that a21 + b21 = 2c21
4
and the above inequality becomes a21 > 7b21. Again from (a) we
obtain
v
u !
b sin b u
1 t a 2 + b2 2
sin 6 CB1M = 1 1
= 1 4a1b1 :
1 1
a1 3 a1 3
p p ;
1
s
sin 6 CB1M = 1p
14x x2 1 < 1 2 49 1 1= 1
7
p p
4 3 4 3
; ; ; ;
which is equivalent to
3x2y + xy2 + 6xy 5x2 y2 24x 3y 27 0:
; ; ; ; ;
and the Sine Law that DB1 = DA sin 6 DAB1 = DA sin 6 DAC =
DA DC (R is the circumradius of ABC ). Analogously DA =
2R 4 1
6
DB DC and DC = DA DB . Our equality is now equivalent to
2R 1
2R
AD CD = BD CD + AD BD and so it su
ces to prove that
(1) DB1 = DA1 + DC1
Denote by m the distance from M to AB and BC and by n the
DM = x (x > 1). Further,
distance from N to AC and BC . Let MN
DB1 = x, DC1 = x 1 and DA1 m = x. Therefore DB = nx,
;
n m ;
n m ;
1
holds.
n-tuples (a1 a2 : : : an) and (b1 b2 : : : bn) formed from distinct ele-
ments of X are called `disjoint' if distinct indices i and j exist such
that ai = bj . Find the maximal number of n-tuples any two of which
are `disjoint'.
ordered n-tuples such that any two of them are `disjoint'. Also let
S (X ) be a set of such n-tuples for which S (X ) = A(n + 1). It is
j j
7
We prove now that A(n + 1) = (n +2 1)! by constructing a set of
(n + 1)! ordered n-tuples, any two of which are `disjoint'. We may
2
assume that X = 1 2 : : : n + 1 . Consider a set E of all even
f g
8
XLVII National
Mathematics Olympiad:
3rd round, 25{26 April 1998
Problem 1. Find the least positive integer number n (n 3)
with the following property: for any colouring of n dierent points
A1 A2 : : : An on a line and such that A1A2 = A2A3 = : : : =
An;1An in two colours, there are three points Ai Aj A2j;i(1 i <
2j ; i n) which have the same colour.
Solution: Assume the two colours are white and black. Consider
8 points coloured as follows: A1 A2 A5 A6 (white), A3 A4 A7 A8
(black). Obviously no three points Ai Aj A2j;i (1 i < 2j ; i n)
have the same colour and therefore n 9.
If we can show that n = 9 has the required property, we will be
done. Suppose there are 9 points coloured black or white and no
three points Ai Aj A2j;i (1 i < 2j ; i n) have the same colour.
First assume that for i = 3, i = 4 or i = 5 points Ai and Ai+2
have the same colour (say white). Then the points Ai;2 Ai+1 Ai+4
1
should be black (note that i ; 2 1 and i + 4 9), which is a
contradiction.
Suppose now that for i = 3 4 5 the points Ai and Ai+2 have
dierent colours. Without loss of generality assume A5 is a white
point. Then A3 and A7 are black. Because of the symmetry we may
suppose that A4 is white and A6 is black. Consiquently A8 is white,
A2 is black (2 + 8 = 2 5) and A9 is white (7 + 9 = 2 8). Therefore
A1 should be both white (1 + 3 = 2 2) and black (1 + 9 = 2 5),
which is again a contradiction.
Consequently the assumption is not true and so n = 9.
2
Note: The assertion is also true if the condition 6 ABC < 90
is left out.
4
1 (a(b2 + c2 ; a2) + b(c2 + a2 ; b2) + c(a2 + b2 ; c2) ; 2abc) =
= 2abc
= cos 6 CAB + cos 6 ABC + cos 6 BCA ; 1:
b.) Applying continuously the equality from a.) and using the
fact that ABCDE is inscribed in a circle convex pentagon, it is easy
to see that
rABC + rAEC + rEDC = rAED + rABD + rBCD :
The condition of the problem and the above equality imply rBCD =
rEDC . Since 4BCD and 4EDC have a side in common and equal
circumradiuses we get that 4BCD = 4EDC (prove it using a.)).
In particular BC = ED and using rABC = rAED again, we get
4ABC = 4AED.
5
(u2 + w2)(v2 + t2) = 2u2v2. Further, since (x1 y1 z1) = 1, it
follows that (u w) = 1 (v t) = 1. Therefore
u2 + w2 = v2 v2 + t2 = 2u2
or
u2 + w2 = 2v2 v2 + t2 = u2:
Without loss of generality we may assume that the rst pair of
equalities hold. It is easily seen that v and u are odd integer numbers.
It follows now from u2 + w2 = v2 that u = m2 ; n2 w = 2mn v =
m2 + n2, where m and n are coprime natural numbers (of distinct
parity). Substitution in v2 + t2 = 2u2 shows that t2 + (2mn)2 =
(m2 ; n2)2 and so t = p2 ; q2 mn = pq m2 ; n2 = p2 + q2 for some
natural numbers p and q. Therefore
p2 q2 = m2(m2 ; p2 ; q2)
which shows that p, q, m is a solution of the original equation. It re-
mains to be seen that pq
m = n < d ( p 2
; q 2
)2 mn = xy , which contra-
z
dicts the way we have chosen x y z.
6
XLVII National
Mathematics Olympiad: 4th
round, 16{17 May 1998
Problem 1. Let n be a natural number. Find the least natural
number k for which there exist k sequences of 0's and 1's of length
2n + 2 with the following property: any sequence of 0's and 1's of
length 2n + 2 coincides in at least n + 2 positions with some of these
k sequences.
; 2
x )Pn;1 (y + 2 x)
!
; 2) + ( (y + x ; 1)( + 1)x P n;1 (y x + 2)
+ (y y
x ) ; 2) n;1 (
x P y x
+ ( + ; 1)( + 1)( ;
2
x y y) n;1 ( + 2 )
x x P y x
+( ; )( + ; 1)( + 1) n;1 ( + 2 )
2
y y x y x P x y :
Solution: Let the square, the n-gon and the m-gon be constructed
on the sides AB , BC and CA, respectively. Denote their centres
by O1, O2 and O3 denote by A1, B1 and C1 the centres of the
equilateral triangles constructed externally on BC , CA and AB .
2
The lines O1C1 O2A1 and O3B1 intersect at the circumcentre O of
4ABC . Since 4A1A2A3 is equilateral, it follows straightforwardly
that 4O1O2O3 is equilateral if and only if C1A1jjO1O2 A1B1jjO2O3
and B1C1jjO1O3. This is equivalent to
OC1 = OA1 = OB1 = k:
C1O1 A1O2 B1O3
On the other hand,
OC1 = cot C + tan 30
C 1 O1 cot 45 ; tan 30
OA1 = cot A + tan 30
A1O2 cot 180n ; tan 30
OB1 = cot B + tan 30 :
B1O3 cot 180m ; tan 30
180 = y. The above identities imply that
Set cot 180
n = x and cot m
cot A = kx ; kp+ 1 cot B = ky ; kp+ 1 cot C = k ; kp+ 1 :
3 3 3
From the identity cot A cot B + cot B cot C + cot C cot A = 1 we get
p
k = p 2(
p x + y + 1 ; 3) p
3xy + ( 3 ; 2)(x + y) p + 3;2
cot C = p x +py ; xy + 3 ; 2 3p :
3xy + ( 3 ; 2)(x + y) + 3 ; 2
p p
Since m n 6 it follows that x y 3,i. e. xy p3(x + y) ; 3.
The inequality cot C p0 implies p x + y ; xy + 3 ; 2 3 0. p and
therefore x + y + 3 ; 2 3 pxy 3(x + y) ; 3 ,i. e. x + y 2 3.
This shows that x = y = 3, i. e., m = n = 6. Hence cot C =
0 cot A = cot B = 1, so 6 C = 90 6 A = 6 B = 45 .
3
Problem 4. Let a1 a2 : : : an be real numbers, not all of them
zero. Prove that the equation
p p p
1 + a1x + 1 + a2x + : : : + 1 + an x = n
has at most one nonzero real root.
X n
Solution: The given equation is equivalent to x p1 +ai a x = n.
i=1 i
a
Since the function p1 + a x (ai =
i
6 0) is strictly decreasing, it follows
i
that this equation has at most one nonzero root.
4
(i) for each colour a and any two vertices A and B of X , the
segment AB is coloured in colour a or there is a vertex C such
that AC and BC are coloured in colour a
(ii) the sides of any triangle with vertices among the vertices of X
are coloured in at most two colours.
Prove that k 2.
5
XLVIII National
Mathematics Olympiad:
3rd round, 17{18 April 1999
Problem 1. Find all triples (x y z) of natural numbers such that
y is a prime number, y and 3 do not divide z, and x3 ; y3 = z2.
Nikolay Nikolov
2
Similar equalities hold for S4NEP S4PEQ and S4QEM . By combin-
ing the above we obtain
SMNPQ = AC BD sin 4AEB (R2 ; OE 2 ) SABCD
6
R2 2
Q. E. D.
3
!
follows that
X
8
ai = 24. It is easy to prove now that
X
8
ai
30.
i=2 ! i =2 2
Therefore 56 =
X8
ai 58, which is false.
i=1 2
0 0 0 0 0 0 0
0 1 1 1 1 0 0
The diagram on the right shows 0 1 1 0 0 1 1
0 0 0 1 1 1 1
that it is possible to have exactly 1 0 1 0 1 0 1
7 contestants: 1 0 1 1 0 1 0
1 1 0 0 1 1 0
1 1 0 1 0 0 1
4
Solution: Note that the incircles of 4ABD and 4ACD are tan-
gent if and only if AB + AD ; BD = AC + AD ; CD, so AB + CD =
AC + BD.
Suppose AB1DC1 is circumscribed and the incircle touches AB1,
B1D, DC1 , C1A in the points M N P Q, respectively. Therefore
AB + CD = AQ + BQ + CP ; DP = AM + BN + CM ; DN =
AC + BD.
Conversely, let the incircles of 4ABD and 4ACD be tangent.
Denote the point of intersection of the tangent through C (dierent
from CA) with the incircle of 4ABB1 by D0 . It follows from the
above that BD0 ; CD0 = AB ; AC = BD ; CD, so DD0 = jCD ;
CD0j. Therefore D0 D, which completes the proof.
Solution: Divide each side of the triangle into ve equal parts
and draw lines parallel to the sides through these points. Thus the
triangle is divided into 25 equilateral triangles of side 51 . The total
number of vertices is 21 > 6 3. Therefore there exist 4 points which
are interior
p to one and the same circle. It is easy to see now that
r 103 , which solves the problem.
5
XLVIII National
Mathematics Olympiad: 4th
round, 18{19 May 1999
Problem 1. The faces of an orthogonal parallelepiped whose di-
mensions are natural numbers are painted green. The parallelepiped
is partitioned into unit cubes by planes parallel to its faces. Find
the dimensions of the parallelepiped if the number of cubes having
no green face is one third of the total number of cubes.
Sava Grozdev
(x ; z)2 +(y ; t)2. It follows from the conditions of the problem that
q = Rr = abc
4S S p = abc(a + b + c)
8S 2
where S is the p p Since S is2 rational (prove it!), it
p area ofpthe triangle.
follows that p ABC ( pA + B + C ) =p8S q is a rational number.
Therefore Ap BC + B AC = 8S 2q ; C AB and after squaring we
obtain that AB is a rational number. Thus AB is a perfect square.
By analogy both BC and CA are perfect squares. Let AB = E 2,
BC = F 2 and CA = G2, where E , F and G are integer. Write A B
and C in the following form: A = a1a22 B = b1b22 and C = c1c22,
where a1 b1 and c1 are square-free integers. So a1b1(a2b2)2 = m2
and therefore a1b1 is a perfect square, whence a1 = b1. By analogy
3
a1 = c1. Thus A = ma21 B = mb21 C = mc21, where m is square-free.
It follows from ma21 = n that m = n a1 = 1 and we obtain
x2 + y 2 = n
z2 + t2 = nb21
(x ; z)2 + (y ; t)2 = nc21
Since both b1 and c1 are integer, it follows from the Triangle Inequal-
ity that 1+b1 > c1 and 1+c1 > b1, whence b1 = c1. It is easy to deter-
mine now that x2 + y2 = 2(xz + yt) and consequently 2(xz + yt) = n.
Let 2(xt ; yz) = k. Then n2 + k2 = 4(x2 + y2)(z2 + t2) = 4n2b21,
so k2 = n2(4b21 ; 1). Therefore 4b21 ; 1 is a perfect square, which is
impossible.
5
;! of 4ABC
troid ;! and with ;! ;2 is 4A0B0C;!
coecient ;! 0 . Therefore
;!
MH0 = ;2 MH and since MH = ;2 MO, we obtain OH0 = ; OH .
6
Union of Bulgarian Mathematicians
BULGARIAN
MATHEMATICAL COMPETITIONS
2000
Soa, 2000
Winter Mathematical Competition
Russe, 4-6 February 2000
Problem 8.1 Given the inequality (n2 ; 1)x < ;3n3 ; 4n2 + n + 2
where n is an integer.
a) Factorize the expression ;3n3 ; 4n2 + n + 2:
b) Find all n for which the inequality holds true for any positive
number x:
Solution: a) ;3n3 ; 4n2 + n + 2 = (n + 1)2(2 ; 3n):
b) Since 0:x < 0 is not true for any x it follows that n 6= ;1: If
n > ;1 then the inequality is equivalent to (n;1)x < (n+1)(2;3n):
If n = 1 then 0:x < ;2 which is not true for any x: Let n 6= 1:
If n ; 1 > 0 then (n + n1)(2 ; 3n) < 0 and the inequality is not
;1
true for any positive x: If n ; 1 < 0 then the inequality becomes
x > (n + n1)(2 ; 3n) and n = 0 is the only solution. If n < ;1 then
;1
(n + 1)(2 ; 3n) > 0 and there exists x > 0 which is not a solution.
n;1
Therefore the inequality has no solution when n < ;1:
Problem 8.2 In an isosceles 4ABC (AC = BC ) the points A1 B1
and C1 are midpoints of BC AC and AB respectively. Points A2
and B2 are symmetric points of A1 and B1 with respect to AB: Let
M be the intersecting point of CA2 and A1C1 and let N be the
intersecting point of CB2 and B1C1: The intersecting point of AN
and BM is denoted by P: Prove that AP = BP:
2
Solution: Since CC1 k A1A2 and CC1 = A1A2 we have that
CC1A2A1 is a parallelogram. Thus, A1M = C1M: But A1B1C1B is
also a parallelogram and therefore the intersecting point of BM and
AC is B1: Hence P lies on the median BB1: Analogously P lies on
the median AA1: In the isosceles 4ABC the medians AA1 and BB1
are of the same length. Therefore AP = 32 AA1 = 23 BB1 = BP:
Problem 8.3 Find all pairs of prime numbers p and q such that
p2 + 3pq + q2 is:
a) a perfect square
b) a power of 5.
Solution: a) Let p2 + 3pq + q2 = r2 where p and q are prime
numbers. If p 6= 3 q 6= 3 then p2 + 3pq + q2 2(mod 3) and
r2 2(mod 3) a contradiction. Without loss of generality p = 3
and we get that q2 + 9q + 9 = r2 and 4q2 + 36q + 36 = (2r)2:
Therefore (2q ; 2r +9)(2q +2r +9) = 45: We may assume that r > 0
and so 2q +2r +9 = 15 or 2q +2r +9 = 45: In the rst case q + r = 3
which is impossible and in the second case solving the system
q+r = 18
2q ; 2r + 9 = 1
we nd q = 7: Because of the symmetry the only solutions are p =
3 q = 7 and p = 7 q = 3:
b) Let p2 + 3pq + q2 = 5n where n is a natural number. Since
p 2 q 2 we have p2 + 3pq + q2 20 and so n 2: It follows
now that 25=(p2 + 3pq + q2) and 5=(p2 + 3pq + q2) = (p ; q)2 + 5pq:
Thus, 5=(p ; q)2 and 25=(p ; q)2: Therefore 25=5pq showing that
p = 5 or q = 5: But if p = 5 then q = 5 (and vice versa). We
obtain p2 + 3pq + q2 = 125 = 53: The only solution of the problem
is p = q = 5:
3
Problem 9a.1 Given the equation jx ;1 2j = jx ;152aj where a is
a parameter.
a) Solve the equation.
b) If a is the square of a prime number prove that the equation
has a solution which is a compose integer.
Solution: After squaring and simple calculations we obtain (26a ;
1)x = (26a ; 1)(26a + 1):
a) If a = 261 then every x 6= 2 is a solution. If a 6= 1 then the
26
only solution is x = 26a + 1:
b) Let a = p2 where p is a prime number. If p = 3 then x = 235
is not a prime. If p 6= 3 then p = 3k 1 and x = 26(3k 1)2 + 1 =
3A + 27 which is divisible by 3:
Problem 9a.2 The quadrilateral ABCD is inscribed in a circle with
diameter BD: Let M be the symmetric point of A with respect to
BD and let N be the intersecting point of the straight lines AM and
BD: The line passing through N which is parallel to AC intersects
CD and BC in P and Q respectively. Prove that the points P C Q
and M are vertices of a rectangle.
Solution: It follows from the condition of the problem that _M lies
on the circumcircle of ABCD: Since <) MAC =<) MBC = MC 2 and
<) MNQ =<) MAC i.e. <) MNQ =<) MBC , we get that the points
M N B and Q lie on a circle. Since <) MNB = 90 we conclude that
<) BQM = 90 : Also, since 4BDC is a right angle triangle we _ have
that MQ k PC: From the other hand <) MDC =<) MAC = 2 and MC
therefore <) MDC =<) MNQ so the points N P M and D lie on a
circle. Thus, <) MPD =<) MND = 90 and MP k CQ: Therefore
P C Q and M are vertices of a rectangle.
4
Problem 9a.3 See Problem 8.3.
Problem 9b.1 Given the system:
1 +x = a;1
x+y
x
x + y = a ; 2
where a is a real parameter.
a) Solve the system if a = 0:
b) Find all values of a such that the system has an unique solu-
tion.
c) If a 2 (2 3) and (x y) is a solution of the system nd all values
of a such that the expression x + y takes its minimal value.
y x
Solution: It follows easily that x and x +1 y are roots of the equation
t2 ; (a ; 1)t + a ; 2 = 0: There are two cases to be considered:
x = 1 x = 1
(1) 1 = a ; 2 () 3 ; a when a 6= 2
x+y y = a ; 2
and
x = a;2 x = a;2
(2) 1 = 1 ()
x+y y = 3 ; a:
(1) AM = BN
AD BD
Further if CH ? AB (H 2 AB ) then 4ALM 4AHC and so
AL = AM : In the same manner BL = BN : But CL is a bisec-
AC AH BC BH
tor and therefore AL = BL so AM = BN : The last equation
AC BC AH BH
combined with (1) gives D H which implies CP ? AB:
Problem 9b.3 Prove that the digit of the hundreds of 21999 +22000 +
6
22001 is even.
Solution: Write the number 21999 +22000 +22001 in the form 21999(1+
2 + 4) = 7:29:21990 = 7:29:210:21980 = 7:29:210:(220)99: Since 29 =
512 210 = 1024 and 220 = (210)2 we have that the last two digits of
220 coincide with the last two digits of 242 so the last two digits of
220 are 76: Moreover the last two digits of 76:76 are also 76: Therefore
the last two digits of the given number are the last two digits of the
product 7:12:24:76 which are 1 and 6: Since 21999 + 22000 + 22001 is
divisible by 8 and it ends by 16 the digit of the hundreds is even.
Problem 10.1 Find all values of the real parameter a such that the
nonnegative solutions of the equation (2a ; 1) sin x +(2 ; a) sin 2x =
sin 3x form an innite arithmetic progression.
Solution: Since sin 2x = 2 sin x cos x and sin 3x = sin x(4 cos2 x;1)
we may write the equation in the form
sin x (2 cos2 x ; (2 ; a) cos x ; a) = 0: Thus sin x = 0 cos x = 1 or
cos x = ; a2 : The nonnegative solutions of the equations sin x = 0
and cos x = 1 are x = k and x = 2k k = 0 1 2a : : : respectively.
Let jaj > 2: In this case the equation cos x = ; 2 has no solution
and therefore the nonnegative solutions of the initial equation are
0 2 : : : which form an arithmetic progression. Let now jaj
2 and let x0 be the only solution of the equation cos x = ; a2 in
the interval 0 ]: In this case the nonnegative solutions of the last
equation are x = x0 + 2k and x = 2 ; x0 + 2k: It is clear now
that the nonnegative solutions form an arithmetic progression only
when x0 = 0 x0 = 2 and x0 = so giving a = ;2 a = 0 and a = 2:
The values of a are a = 0 and jaj > 2:
Problem 10.2 Let O I and H be respectively the circumcenter,
incenter and orthocenter for an acute nonequilateral 4ABC: Prove
7
that if the circumcircle of 4OIH passes through one of the vertices
of 4ABC then it passes through another vertex of 4ABC:
Solution: Assume that O I H and C lie on a circle. It is well known
that CI is the bisector of <) HCO: Thus <) IHO =<) ICO =<) ICH
=<) HOI and it follows from 4IHO that IH = IO = t:
We shall prove that if <) BAC 6= 60 then O I H and A lie on
a circle. Denote by M and N the projection points of I respec-
tively on AO and AH: Let O1 and O2 be such that IO1 = IO2 = t
and O1 lies between A and M and M lies between O1 and O2:
Analogously let H1 and H2 be such that IH1 = IH2 = t and H1
lies between A and N and N lies between H1 and H2: If O
O1 H H1 or O O2 H H2 then 4AIO 4AIH and there-
fore AO = AH: But AH = 2AO cos <) BAC and so <) BAC = 60 :
If O O1 H H2 or O O2 H H1 then it follows from
<) IO1O2 =<) IO2O1 =<) IH1H2 =<) IH2H1 that AOIH is inscribed.
Suppose now that A and B do not lie on the circumcircle of
4OIH: In this case <) BAC =<) ABC = 60 and therefore 4ABC
is equilateral which is a contradiction.
Problem 10.3 In each of the cells of a 3
3 table is written a real
number. The element in the i;th row and j ;th column equals to
the modulus of the dierence of the sum of the elements from the
i;th row and the sum of the elements from the j ;th column. Prove
that every element of the table equals either to the sum or to the
dierence of two other elements of the table.
Solution: Let p1 p2 p3 and q1 q2 q3 be the sum of the elements in
the rst, second and third row and in the rst, second and third
column respectively. It is clear that p1 + p2 + p3 = q1 + q2 + q3:
Therefore the element in the rst row and the rst column equals to
jp1 ; q1j: From the other hand jp1 ; q1j = jp2 + p3 ; q2 ; q3j which
implies jp1 ; q1j = 1jp2 ; q2j + 2jp3 ; q3j where 1 2 2 +1 ;1: Since
8
jp1 ; q1j 0 it is clear that 1 = 2 = ;1 is impossible. Therefore
jp1 ; q1j is either the sum or the dierence of jp2 ; q2j and jp3 ; q3j:
By analogy every element of the table is the sum or the dierence
of two other elements.
Problem 11.1. Prove that for every positive number a the se-
quence fxng1 n=1 such that x1 = 1 x2 = a xn+2 = xn+1 xn n 1
q
3 2
n k
1) = ; 31 : Adding the equalities k+2 ; k+1 = ; 13 for k =
n+1
1 ; ; 13
1 1
1 n 3 (1 ; ; 1 n+1 ): Since
;3 + : : : + ;3 = = 4 3
1 + 13
0 = 0 and nlim
p2 = 4a + 129b ; 3c q2 = 3b + 69c ; 2a
2 2 2 2 2 2
(2)
b) It follows from the condition of the problem that <) ADB >
90 and so c > a: Let D1 be a point on AB such that D1B = DB:
Further <) AD1B =<) ABD + 180 ; 2<) ABD = 180 + 2<) ABD =
<) ADB: Therefore 4AD1D 4ADB and thus b2 = c(c ; a): A
substitution in (2) implies p = 3c ;3 2a and q2 = 9c ; 3ac ; 2a2 :
2
9
Hence q2 = p(p ; a): Let B1 be a point on BC (B lies between C
and B1), such that BB1 = BD = a: Since qp = p +q a it follows that
4DBC 4B1DC: Therefore <) DBC =<) BB1D+ <) BDB1 =
2 <) BB1D = 2 <) BDC:
Problem 11.3 See problem 10.3.
10
Spring Mathematical Tournament
Jambol, 24-26 March 2000
Problem 8.1. Let f (x) be a linear function such that f (0) = ;5
and f (f (0)) = ;15: Find all values of m for which the set of the
solutions of the inequality f (x):f (m ; x) > 0 is an interval of length
2:
Solution: Let f (x) = ax + b: It follows from f (0) = ;5 that b = ;5
and from f (f (0)) = ;15 that a = 2: Therefore the function is f (x) =
2x ; 5: Consider the inequality (2x ; 5)(2(m ; x) ; 5) > 0 ()
(2x ; 5)(2m ; 5 ; 2x) > 0: The solution of the last inequality is an
interval with end points 25 and 2m2; 5 : Therefore 25 ; 2m2; 5 = 2
so j5 ; mj = 2: Finally, we obtain that m = 3 and m = 7:
Problem 8.2 Given an isosceles right angle triangle ABC with
<) ACB = 90 : Point P lies on BC M is the midpoint of AB and
let L and N be points from the segment AP such that CN ? AP
and AL = CN:
a) Find <) LMN
b) If the area of 4ABC is 4 times greater than the area of
4LMN nd <) CAP:
Solution: a) Let <) CAP = : We nd <) ACN = 90 ; and
<) MCN =<) ACN ; 45 = 45 ; =<) LAM: Since AM = CM and
AL = CN it follows that 4AML = 4CMN: Therefore <) AML =
<) CMN and so <) LMN = 90 ; <) AML+ <) CMN = 90 :
b) It follows from 4AML
= 4CMN that LM = MN and since
11
<) LMN = 90 we get SLMN = MN Also SABC = AC : Now
2 2
2
: 2
4MN 2 = AC 2 applies that MN = AC : Denote by Q the midpoint
2 2 2
AC
of AC: Thus QM = QN = 2 = MN: Therefore 4QMN is
equilateral and <) QNM = 60 which implies <) QNA = 60 ; 45 =
15 : But AQ = QN and so <) CAP =<) QNA = 15 :
Problem 8.3 There are 2000 white balls in a box. There are also
suciently many white, green and red balls. The following opera-
tions are allowed:
1) Replacement of two white balls with a green ball
2) Replacement of two red balls with a green ball
3) Replacement of two green balls with a white bal and a red
ball
4) Replacement of a white ball and a green ball with a red ball
5) Replacement of a green ball and a red ball with a white ball
a) After nitely many of the above opperations there are three
balls left in the box. Prove that at least one of them is a green ball.
b) Is it possible after nitely many opperations to have only one
ball left in the box?
Solution: Consider a box with x white, y green and z red balls.
Direct verication shows that after applying any of the allowed op-
erations the sum x + 2y + 3z does not change modulo 4. Since the
initial values are x = 2000 y = z = 0 we obtain that this sum is
congruent to 0 modulo 4.
a) There are 3 balls in the box and therefore x + y + z = 3:
Moreover x +2y +3z 0(mod4): If a green ball is not in the box then
y = 0 and so x + z = 3 and x +3z x +3(3 ; x) 1 ; 2x 0(mod4)
12
which is impossible.
b) Suppose that there is only one ball left in the box. Therefore
x + y + z = 1 and x + 2y + 3z 0(mod4) which is impossible.
Problem 9a.1 Find all values of m such that the equation
1 + m ; 2m (jx ; mj ; m) = 0
x + m x ; m m2 ; x2
has exactly one nonnegative root.
Solution: The equation x +1 m + x ;mm ; m22;m x2 = 0 is equivalent
to (when x 6= m) to (m + 1)x = ;m(m + 1): When m = ;1 it has
innitely many roots { all numbers x 6= 1: If m 6= ;1 we obtain
x = ;m: The equation jx ; mj ; m = 0 has two roots: x = 0 and
x = 2m when m > 0 an unique root: x = 0 when m = 0 and has
no roots when m < 0:
Let m < 0 and m 6= ;1: In this case the equation has an unique
root x = ;m and it is nonnegative. Let m = 0: Then the equation
becomes x1 jxj = 0 which obviously has no roots. Let m > 0: The
equation has three roots x = ;m x = 0 x = 2m and two of them
are nonnegative.
Thus, the desired values of m are m < 0 and m 6= ;1:
Problem 9a.2 Given an acute-angled triangle ABC and let
(x + y)2 = t2 + 2t1:
equalities t2 ; 2t1 0 and t2 + 2t1 0 hold true. It suces to prove
that when A < 14 the roots t1 < t2 of the equation g(t) = t2 ; t + A
satisfy the inequalities t2 ; 2t1 0 and t2 + 2t1 0: If A = 0
then t1 = 0 and t2 = 1 and the inequalities hold true. It is clear
that we can consider only the case A 2 (0 41 ): Now t1 > 0 t2 > 0
and the inequality t2 + 2t1 0 holds true. From the other hahd
t1 + t2 = 1 and the inequality t2 ; 2t1 0 is equivalent to t2 23 :
The same inequality is equivalent to t1 13 : For the roots t1 and
t2 we obtain t1 31 t2 23 which is equivalent to gg(( 32 )) 0
1
3p 0:
15
the other hand <) RC1A =<) RHA =<) AB1T and therefore B1ATC1
is inscribed. Hence, T = k \ B1Q: This implies that T S which
shows that the three lines intersect in a point.
Problem 9b.3 See Problem 8.3.
Problem 10.1 Solve the equation px + p3 x + 7 = p4 x + 80:
Solution: It is clear that x 0: Note that x = 1 is a root of
the equation. We shall prove that there are no other p roots.
p Ras-
ingpthe p equation to the p forth
p power wepget x + 4( x) x + 7 +
2 3 3
= x + 80:
Let f ( x ) = 4( px)3 p3 x + 7 + 6(px)2(p3 x + 7)2
p p p
+4( x)( 3 x + 7)3 +( 3 x + 7)4: Obviously f (x) is an increasing func-
tion. If x > 1 then it follows from the inequalities x2 > x f (x) >
f (1) = 80 that x2 + f (x) > x + 80 a contradiction. If x < 1
then it follows from the inequalities x2 < x f (x) < f (1) = 80 that
x2 + f (x) < x + 80 a contradiction. This completes the prove.
Problem 10.2 The incircle of an isosceles 4ABC touches the legs
AC and BC at points M and _
N respectively. A tangent t is drawn
to the smaller of the arcs MN and let t intersects NC and MC at
points P and Q respectively. Let T be the intersecting point of the
lines AP and BQ:
a) Prove that T lies on the segment MN
b) Prove that the sum of the areas of triangles ATQ and BTP
is the smallest possible when t is parallel to AB:
Solution: a) Let the incircle touches AB and PQ at points R and
S respectively. Let MN and SR intersect QB in points T1 and
T2 respectively. Since <) T1MQ =<) T1NC = 180 ; <) T1NB and
<) MT1D =<) BT1N it follows from the Law of Sines for 4MT1D and
16
DT1 = BT1 so QT1 = MQ : By analogy QT2 = SQ :
4BT1N that MD BN BT BN 1 BT BR 2
q s
by 3 and after multiplying 2:AB:AD cos 1(mod 3) by AC 2 we
obtain 2:AC 2:AB:AD: cos cos
1(mod 3) a contradiction to
AC 2:AB:AD: cos cos
1(mod 3): Therefore at least one of the
distances is divisible by 3.
Let n 4: Since there exist n4 sets with four elements each
!
4
by 3. Each such distance is counted exactly n ;2 2 times and we
!
n
= 16 n2 :
!
18
a) Find the greatest value of f (x)
f (x)
x 2
; 5x +
b) Find the greatest value of the function x2 + 5x + 20 10
!
:
Solution: a) We shall prove that the greatest value of f (x) equals
2. Since x2 + 7x + 14 > 0 8x we obtain f (x) 2 =) (x + 5)2 0
and an equality occurs only when x = ;5:
b) Let g(x) = xx2 ; 5x + 10 : Since x2 +5x +20 > 0 8x we obtain
2
+ 5x + 20
g(x) 3 =) (x + 5)2 0 and equality occurs only when x = ;5:
Since x2 ; 5x +10 > 0 8x we have that the function h(x) = g(x)f (x)
is correctly dened. Further, when f (x) > 0 i.e. x 2 ;3 ;1] then
h(x) 32 = 9: From the other hand g(x) 1 =) x ;1 and so
h(x) 1 when f (x) 0: Therefore the greatest value of h(x) equals
9 and h(x) = 9 when x = ;5:
Problem 11.2 A point A1 is chosen on the side BC of a triangle
ABC such that the inradii of 4ABA1 and 4ACA1 are equal. De-
note the diameters of the incircles of 4ABA1 and 4ACA1 by da:
In the same manner dene db dc: If BC = a CA = b AB = c p =
a + b + c and h h h are the altitudes of the triangle ABC and d
a b c
2
is the diameter of the incircle of 4ABC prove that:
q
p(p ; a)
a) da + a d = ha
b) da + db + dc + p ha + hb + hc :
Solution: a) Let O1 and O2 be the incenters of 4ABA1 and 4ACA1
and p1 and p2 be semiperimeters of the same triangles. We have
SABC = SABA1 + SACA1 = ra:p1 + ra:p2 = rc(p1 + p2) = rc :(p + CC1)
where ra = d2a : Therefore ra(p + AA1) = SABC : If P and Q are
the touching points of the incircles of 4ABA1 and 4ACA1 with
19
BC then the quadrilateral O1PQO2 is rectangle and so O1O2 =
PQ = PA1 + QA1 = p1 ; AB + p2 ; AC = p1 + p2 ; AC ; BC =
p + AA1 ; AC ; BC: It follows from the similarity of 4IO1O2 and
4IBC where I is the incenter of 4ABC that OBC 1 O2
= r ;r ra =)
AA1 + p ; AC ; BC = r ; rc : We obtain the system
BC r
ra(p + AA1) = S q
AA1 + AB ; p = r ; ra Thus, AA1 = p(p ; a) and ra =
c p pr p q
rp =
r p( p ; p ; a) = ha ; r p(p ; a) which
a 2 a
q
p + p(p ; a)
implies a).
pp ; a pp ; b
b) By a) the inequality is equivalent to a + b +
pp ; c pp pp ; a pp ; b pp ; c
c p d : Further p a + b + p c =
p;a p;b p;c
(p ;pb) + (p ; c) (p ;p a) + (p ; c) (pp; a) + (p ; b)
+ +
p p ;pa + p p ;pb + p p ;pc =
2 p ; b p ; c 2 p ; ap p ; c 2 p ; a p ; b
p p
q = d:
(p ; a)(p ; b)(p ; c)
Problem 11.3 See problem 10.3.
20
XLIX National Mathematical Olympiad
Third Round, 15-16 April 2000
Problem 1. Find all value of the real parameter a such that the
equation
9t ; 4a3t + 4 ; a2 = 0
has an unique root in the interval (0 1):
Solution: After the substitutions x = 3t and f (x) = x2 ; 4ax +
4 ; a2 the problem is equivalent to: nd all values of a such that
the equation f (x) = 0 has an unique root in the interval (1 3): It
is easily seen that if f (1) = 0 or f (3) = 0 then a = 1 a = ;5 or
a = ;13 are not solutions of the problem. Therefore the equation
f (x) = 0 has an unique root in the interval (1 3) when f (1):f (3) < 0
or Df = 0 2a 2 (1 3): It follows now that f (1):f (3) < 0 () (a ;
p < 0 () a 2 (;13 ;5) or 5a2 ; 4 = 0 2a 2 (1 3)
1)2(a +5)(a +13)
and so a = 2 5 5 :
22
Solution: Let <) AEB = 90 : Since <) CEB < 90 there exists a
point F on the side AB such that <) BEF =<) BEC: Thus 4BCE
=
4BFE which implies BF = BC and <) BFE =<) BCE: From the
other hand
() sin <AE = AF AE = AD :
) AFE sin <) AEF sin <) ADE sin <) AED
Since <) AED =<) AEF and <) AFE + <) ADE = 180 we obtain
AF = AD and so AB = AD + BC:
Conversely, let AB = AD + BC: There exists a point F on
the segment AB such that AF = AD and BF = BC: Therefore
4BFE = 4BCE so <) BFE =<) BCE and <) BEF =<) BEC: It
follows from () and AF = AD that sin <) AED = sin <) AEF: Since
<) AED+ <) AEF < 180 we have <) AED =<) AEF and therefore
<) AEB = 90 :
Problem 5. Prove that for any two real numbers a and b there
exists a real number c 2 (0 1) such that
ac + b + c +1 1 > 241 :
25
XLIX National Mathematical Olympiad
Fourth Round, 16-17 May 2000
Problem 1. In an orthogonal coordinate system xOy a set consisit-
ing of 2000 points Mi (xi yi) is called "good" if 0 xi 83 0
yi 1 i = 1 2 : : : 2000 and xi 6= xj for i 6= j: Find all natural
numbers n with the following properties: : a) For any "good" set
some n of its points lie in a square of side length 1:
b) There exists a "good" set such that no n + 1 of its points lie
in a square of side length 1:
(A point on a side of a square lies in the square).
Solution: We shall prove that n = 25 is the only solution of the
problem. We show rst that for a "good" set some 25 points lie
in a square of side length 1. All points lie in the rectangle 1
x 83 0 y 1: Divide this rectangle to 83 squares of side
length 1: If some 25 points lie in one of these rectangles then we
are done. Conversely, in every square there are at least 26 or at
most 24 points. We prove now that there exists a square with at
least 26 points and there exists a square with at most 24 points.
Indeed, if a square with 26 points does not exist then the points are
at most 83:24 = 1992 < 2000: If there is no square with less than
25 points then the points are 83:26 ; 82 > 2000: Further, move the
square with more than 25 points towards the square with less that
25 points. Since the number of points in this square changes at most
by one we get the assertion of the problem.
To prove b) let x1 = 0 xi = xi;1 + 1999 83 and y = 0 for i =
i
0 2 4 : : : 2000 while yi = 1 for i = 1 3 : : : 1999: Let XY ZT be an
unit square. WLOG we assume that it intersects the lines
26
y = 0 and y = 1 in the points P Q and R S respectively. Then
PQ+QR+RS +SP < PY +Y Q+QZ +ZR+RT +TS +SX +XP = 4:
But SP QR 1 and we get PQ + RS < 2: This implies that in the
square XY ZT there are no more than 25 points.
We found that n = 25 is a solution of the problem. It is unique
since 24 < 2000
83 < 25:
Problem 2. Given an acute 4ABC: Prove that there exist unique
points A1 B1 and C1 on BC AC and AB respectively with the fol-
lowing property: Each of the points is the midpoint of the segment
with ends the orthogonal projections of the other two points on the
corresponding side. Prove that 4A1B1C1 is similar to the triangle
formed by the medians of 4ABC:
Solution: It is easy to be seen that the perpendicular through C1 to
AB intersects A1B1 at its midpoint. The same is applied for A1 and
B1: Therefore the tree perpendiculars intersect at the medcenter of
4A1B1C1; the point P:
27
Let T be the midpoint of A1B1 <) PB1A1 = x =<) PCA1 <) PA1B1 =
y =<) PCB1: The Law of Sines for 4B1TP and 4A1TP gives:
sin = B1T = A1T = sin
sin x TP AP sin y
where sin x = sin : Further, if CN is the median in 4ABC similar
sin y sin
<) ACG+ <) BCG = and is acute it follows that x =<) ACG y =
<) BCG: It is clear now that CP is symmetric to the median in
4ABC through C with respect to the bisector of angle : The same
is true for AP and BP: Therefore the point P is unique (and there-
fore A1 B1 and C1 are unique). Further, <) B1C1A1 =<) B1C1P +
<) A1C1P =<) B1AP + <) A1BP =<) BAG+ <) ABG =<) BGK where
K is the symmetric point of G with respect to the midpoint of BC:
By analogy <) C1A1B1 =<) GBK and <) A1B1C1 =<) GKB: There-
fore 4A1B1C1 is similar to the triangle formed by the medians of
4A1B1C1:
28
Problem 3. Let p 3 be a prime number and a1 a2 : : : ap;2 be
a sequence of natural numbers such that p does not divide both ak
and akk ; 1 for all k = 1 2 : : : p ; 2: Prove that the product of some
elements of the sequence is congruent to 2 modulo p:
Solution: Consider the sequence 1 a1 a2 : : : ap;2: We shall
prove by induction that for any i = 2 3 : : : p ; 1 there exist in-
tegers b1 b2 : : : bi each of which is a product of some elements from
the above sequence and bm 6 bn(mod p) for m 6= n: Indeed, for
i = 2 we can choose b1 = 1 b2 = a1(a1 6 1(mod p)): Suppose we
have chosen b1 b2 : : : bi such that bm 6 bn(mod p) for m 6= n: Con-
sider b1ai b2ai : : : biai: It is easily seen that any two of them are not
congruent modulo p: Further, if bj ai for any j is congruent to bl for
some l we get that b1ai b2ai : : : biai modulo p is a permutation of
b1 b2 : : : bi modulo p: Thus,
(b1ai):(b2ai): : : : :(biai) b1:b2: : : ::bi(mod p)
and therefore aii 1(mod p); a contradiction.
It follows now that for any s = 2 3 : : : p ; 1 there exist few
elements of the sequence from the condition of the problem such
that their product is congruent to s modulo p:
F H DO
0 0 F H GG
It follows from GG00 =0 DO0 O0G0 = DG and <) DBG = 2 ; 90
that F 0H 0 = BD:O GD
D = ;tg2 :O0D: Denote BC \ EF 00 = I: Since
<) CBF 00 = 180 ; 2 and BE = 2O0D it is clear that F 00H 00 =
BF 00 sin = cos(180 BI sin = ;BE sin cos = ;O0Dtg2 =
; 2 ) cos 2
F 0H 0: This completes the prove.
Problem 6. Let A be the set of all binary sequences of length n
and let 0 2 A be the sequence with zero elements. The sequence c =
c1 c2 : : : cn is called sum of a = a1 a2 : : : an and b = b1 b2 : : : bn
if ci = 0 when ai = bi and ci = 1 when ai 6= bi: Let f : A ! A be
a function such that f (0) = 0 and if the sequences a and b dier in
exactly k terms then the sequences f (a) and f (b) dier also exactly
in k terms. Prove that if a b and c are sequences from A such that
a + b + c = 0 then f (a) + f (b) + f (c) = 0:
Solution: Consider the sequence e1 = 1 0 : : : 0 0 e2 = 0 1 : : : 0
: : : en;1 = 0 0 : : : 1 0 en = 0 0 : : : 0 1: It follows from the condi-
31
tion of the problem that (since f (0) = 0) for all p 1 p n there
exists q 1 q n such that f (ep) = eq:
It is clear also that f (ep) 6= f (eq ) for 1 p q n p 6= q:
Therefore
(1) ff (e1) f (e2) : : : f (en)g = fe1 e2 : : : eng:
Consider an arbitrary sequence a = a1 a2 : : : an with t ones. If
f (ep) = eq and ap = 1 then the q;th term of the sequence f (a) is
also 1 (otherwise ep and a dier at t ; 1 terms whereas f (ep) = eq
and f (a) dier at t + 1 terms). By analogy if ap = 0 then the q;th
term of f (a) is also 0.
Finaly, consider the sequences a = a1 a2 : : : an b = b1 b2 : : : bn
and c = c1 c2 : : : cn such that a + b + c = 0: This means that
for every i 1 i n the sum ai + bi + ci is even number. Fix
i 1 i n and let f (ei) = ej : It follows now that the j ;th terms of
the sequences f (a) f (b) f (c) coincide with ai bi ci respectively and
using (1) we obtain that f (a) + f (b) + f (c) = 0:
32
Union of Bulgarian Mathematicians
BULGARIAN
MATHEMATICAL COMPETITIONS
2001
Soa, 2001
Winter mathematics competition
Bourgas, 2-4 February 2001
Problem 9.1. a) Draw all points in the plane with coordinates
(x y) such that
(j3x ; yj ; 3)(j3x + yj ; 3) = 0
b) Find all x and y for which
(j3x ; yj ; 3)(j3x + yj ; 3) = 0
y ; f4xg = 0
;1 x 1
(For a real number x we denote the unique number in the interval
0 1) for which x ; fxg is an integer by fxg).
Solution: a) It is easy to see that these are the points on the four
lines l1 : 3x ; y = 3 l2 : 3x ; y = ;3 l3 : 3x + y = 3 and l4 : 3x + y =
;3:
b) There are two types of solutions:
y = ;3x + 3
1. Solutions of y = f4xg
;1 x 1
y = 3x + 3
2. Solutions of y = f4xg
;1 x 1
Denote the integer part of 4x by 4x] i.e. 4x = 4x] + f4xg: In
the rst case we have ;3x + 3 = f4xg = 4x ; 4x] i.e 7x = 3 + 4x]:
3
Since the right hand side is an integer all posiible values of x are
0 71 72 : : : 1: Direct verication shows that only x = 75 y = 67
x = 76 y = 37 x = 1 y = 0 are solutions.
In the second case 3x + 3 = f4xg = 4x ; 4x] i.e. x = 3 + 4x]:
It follows that x is an integer, i.e. x = ;1 0 or 1: Direct verication
shows that solution is only x = ;1 y = 0.
Thus, there are four solutions:
x = 75 y = 76 x = 76 y = 37 x = 1 y = 0 x = ;1 y = 0:
Problem 9.2. Points A1 B1 and C1 are chosen on the sides BC CA
and AB of a triangle ABC: Point G is the centroid of 4ABC and
Ga Gb and Gc are centroids of 4AB1C1 4BA1C1 and 4CA1B1
respectively. The centroids of 4A1B1C1 and 4GaGbGc are denoted
by G1 and G2 respectively. Prove that:
a) the points G G1 and G2 lie on a straight line
b) lines AGa BGb and CGc intersect in a point if and only if
AA1 BB1 and CC1 intersect in a point.
Solution: a) Let O be an arbitrary point and
OA~ 1 = OB
~ + (1 ; )OC ~ OB
~ 1 = OA
~ + (1 ; )OC
~
~ 1 = OA
OC ~ + (1 ; )OB
~
where 2 (0 1): The existence of and follows from
the fact that A1 B1 and C1 lie on the sides BC CA and AB of
4ABC: Then we have
~ = 1 OA ~ + OB~ + OC ~ = 1 3OA ~ + 3OB~ + 3OC ~
OG 3 9
OG~ 1 = 13 OA
~ 1 + OB
~ 1 + OC
~1 =
4
1 ( + )OA
h
~ + ( ; + 1)OB ~ + (2 ; ; )OC
~ :
i
3
By analogy OG ~ a = 1 (1 + + )OA
h
~ + (1 ; )OB~ + (1 ; )OC
~ i
3
~ b = 1 OA~ + (2 + ; )OB
~ + (1 ; )OC ~
h i
OG 3
~ c = 1 OA~ + OB ~ + (3 ; ; )OC
~ :
h i
OG 3
Thus, OG~ 2 = 1 OA
~ 1 + OB
~ 1 + OC
~1 =
3
1 (1 + 2 + 2 )OA
h
~ + (3 + 2 ; 2 )OB
~ + (5 ; 2 ; 2 )OC~ : i
9
Since GG~ 1 = OG~ 1 ; OG~ we have
~ 1 = 1 ( + ; 1)OA
GG
h
~ + ( ; )OB
~ + (1 ; ; )OC
~ : i
3
Using the same arguments we obtain GG ~ 2 = OG
~ 2 ; OG
~ =
1 (2 + 2 ; 2)OA
h
~ + (2 ; 2 )OB~ + (2 ; 2 ; 2 )OC~ : It follows
i
9
~ 1 = 3 GG
from the last two equalities that GG ~
2 2 and we are done.
b) Since the lines AA1 BB1 and CC1 intersect in a point Ceva's
theorem gives AC1 BA1 CB1 = 1:
C1B A1C B1A
Denote the intersecting points of AGa BGb and CGc with the sides
BC CA and AB by A2 B2 and C2 respectively. A necessary and
sucient condition for lines AGa BGb and CGc to intersect in a
CA2 BC2 AB2 = 1:
point is A
2 B C2A B2 C
Denote the midpoint of B1C1 by A3: Let h1 and h2 be the alti-
5
tudes of 4AA2C and AA3B1 from A2 and A3: We have
SAA2 C = h1:AC = AA2:AC :
SAA3 B1 h2:AB1 AA3:AB1
By analogy
SAA2 B = AA2:AB :
S AA3 C1AA :AC 3 1
Dividing the above equalities and using that SAA3 C1 = SAA3 B1 and
SAA2 C = CA2 we obtain
SAA2 B BA 2
CA2 = AC AC1 :
A2B AB AB1
By analogy
BC2 = CB CB1 AB2 = BA BA1 :
C2A CA CA1 B2C BC BC1
Multiplying the above equalities gives
CA2 BC2 AB2 = AC AC1 BC CB1 BA BA1 = AC1 BA1 CB1 :
A2B C2A B2C AB AB1 AC CA1 BC BC1 C1B A1C B1A
Therefore the lines AGa BGb and CGc intersect in a point i the
lines AA1 BB1 and CC1 interect in a point.
Problem 9.3. Let An be the number of sequences from 0's and 1's
of length n such that no four consequtive elements equal 0101: Find
the parity of A2001:
Solution: Denote the number of sequences of length n, the last
three terms of which are ijk where i j k 2 f0 1g and no four con-
sequtive elements equal to 0101 by anijk :
6
Obviously anijk+1 = an0ij + an1ij when ijk 6= 101 and an101+1 = an110:
Adding the above equalities for all values of ijk we obtain
An+1 = 2An ; an010:
Therefore An+1 an010 an001;1 + an101;1 an000;2 + an100;2 + an110;2
a0003 + an100;3 a000
n; n;4 + an;4 + an;4 + an;4 an;5 + an;5 + an;5 +
100 010 110 000 001 010
an100;5 + an111;5 + an110;5 + an101;5 + an011;5+ An;5(mod 2):
Thus, Ak+6m Ak (mod 2) for all k and m:
Therefore A2001 A3 = 8 0(mod 2):
Problem 10a.1. Find all pairs (a b) of integers such that the system
x2 + 2ax ; 3a ; 1 = 0
y2 ; 2by + x = 0
has exactly three real solutions.
Solution: Let (a b) satisfy the condition of the problem. Since there
are exactly three real solutions the rst equation has two distinct
roots x1 < x2 which implies D1 = 4(a2 + 3a + 1) > 0: The discrim-
inant of the second equation in respect to y equals D2 = 4(b2 ; x)
and therefore there are exactly three real solutions i x1 < x2 = b2:
These conditions are satised pi a and b are integers such that
a + 3a + 1 > 0 and b2 = ;a + a2 + 3a + 1: It follows from the
2
7
Problem 10a.2. The tangential point of a circle k through the
vertex C of a 4ABC and the line AB is the vertex B . The circle k
intersects for a second time the side AC snd the median of 4ABC
through C at points D and E respectively. Prove that if the inter-
secting point of the tangents to k through C and E lies on the line
BD then <) ABC = 90:
Solution: Let F be the intersecting point of the tangents tC and
tE to k at C and E G = BD \ CE and ACBH is parallelogram.
We have that <) ABC = 90 () tC k AB () FD FB CA= CD :
CD CD GD
From the other hand CA = BH = GB i.e. it suces to show that
FD = GD (
): Since 4FBE 4FED and 4FBC 4FCD
FB GB
we obtain FBFE ED= BE and FB = BC : Further, FC 2 = FE 2 =
FC CD
FD CD:ED
FB:FD and so FB = CB:EB : Now the equality from (
) follows
from CD:ED = SCED = GD which completes the proof.
CB:EB SCEB GB
Problem 10a.3. Ivan and Peter alternatively write down 0 or 1
until each of them has written 2001 digits. Peter is a winner if the
number, which binary representation has been obtained, cannot be
expressed as a sum of two perfect squares. Prove that Peter has a
winning strategy.
Solution: First we prove that if the binary representation of a pos-
itive integer ends with two ones and even number of zeroes then
this integer cannot be represented as sum of two squares. Indeed,
such a number is of the form 4k (4s + 3) and if we suppose that
x2 + y2 = 4k (4s + 3) then using the fact that x2 + y2 equals 0 mod-
ulo 4 i x and y are even we get that there exist integers p q for
which p2 + q2 = 4s + 3 which is a contradiction.
8
The winning strategy of Peter could be:
If one of Ivan's digits is 1 then Peter simply repeats all digits
written by Ivan. The nal number is of the form 4k (4s + 3) and
cannot be written as x2 + y2: If all Ivan's digits are zeroes then
the rst three digits of Peter are 1 1 1 after which he writes only
zeroes. The nal number is (0101010 : : : 0)2 = 21:41998 and cannot
be represented as p2 + q2 since 21 cannot be written in that form.
Problem 10b.1. Find all values of the real parameter a such that
the equation
logx (x2 + x + a)2 = 4
has unique solution.
Solution: The equation is equivalent to (x2 + x + a)2 = x4 x >
0 x 6= 1: Further, (x2 + x + a)2 = x4 () p (x + a)(2x2 + x +
a) = 0 with roots x1 = ;a and x2=3 = ;1 4 1 ; 8a provided
; 1 ; p1 ; 8a
1 ; 8a 0: Since 4 < 0 we obtain that only x1 =
p
;a x2 = ;1 + 4 1 ; 8a can be roots of the equation from the prob-
lem. If a 0 then x1 0 and x2 0 which implies that the
equation has no roots. If a < 0 then x1 > 0 and x2 > 0: Note
that x1 = x2 implies a = 0 which is a contradiction. Therefore if
the equation has unique solution then it is necessary to have a < 0
and one of the two roots equals 1. Thus, x1 = 1 ) a = ;1 and
x2 = 1 ) a = ;3:
Problem 10b.2. On each side of a right isosceles triangle with
legs of length 1 is chosen a point such that the triangle formed from
these three points is a right triangle. What is the least value of the
hypotenuse of this triangle?
9
Solution: Consider a right isosceles triangle ABC with right angle
at C . Let A1 2 BC B1 2 CA and C1 2 AB be such that 4A1B1C1
is right triangle.
1. Suppose <) B1A1C1 = 90 : Assume that the circle k with di-
ameter B1C1 intersects BC at point X 6= A1: It is easy to be seen
that X is an interior point for the line segment BC: Draw a tangent
l to k which is parallel to BC
_
and the tangential point Y belongs
to the smaller of the arcs A1X : Denote the intersecting points of l
with AB and AC by C2 and B2 respectively. Consider a homothety
of center A such that the image of B2(C2) is C (B ): It is clear that
the image of 4B1C1Y is inscribed in 4ABC and its hypotenuse
is less than B1C1: Therefore wlog we may assume that BC is tan-
gent to k: Thus, if <) C1B1A1 = then <) C1A1B = : From the
Sine Low for 4BC1A1 we obtain BA1 = B1C1 sinsin sin( + 45) =
45
B1C1 sin (sin + cos ): Hence
(1) 1 = BA1 + A1C = B1C1 cos sin + B1C1 sin (sin + cos )
Therefore 1 = B1C1(sin2 + sin 2) () B1C1(1 + 2 sin 2 ;
cos 2) = 2: From the other hand 2 sin 2 ; cos 2 =
p 2 p p
5 p sin 2 ; p1 cos 2 = 5 sin(2 ;
) 5 where
is
!
5 5
such that sin
= p and cos
= p2 : Note that the equality holds
1
5 5
when 2 ;
= 90 and since
< 90 we have that < 90:
p
Further Bp1C1(1 + 5) B1C1(1 + 2 sin 2 ; cos 2p) = 2 and
so B1C1 52; 1 : It is easily seen that if B1C1 = 52; 1 and
2 ;
= 90 p then (1) holds true and therefore there exists triangle
with B1C1 = 52; 1 :
2. Suppose <) A1C1B1 = 90: Denote the projection of A1 and
10
B1 on AB by A2 and B2 respectively. Obviously AB2 = B1B2
and BA2 = A1A2: Let P be the midpoint of A1B1 and Q { the
midpoint of A2B2: Thus, A1B1 = 2PC1 2PQ = A1A2 + B1B2 and
A1B1 A2B2 = AB ; AB2 ; BA2 = AB ; B1B2 ; A1A2: Adding p
p
the above equalities gives 2A1B1 AB = 2 and so A1B1 22 :
p p
Since 22 > 52; 1 we obtain that the least possible value equals
p
5 ; 1:
2
Problem 10b.3. An element x is chosen from the set
A = f1 2 : : : 2n g n 3: Questions of the type: Does x belong
to B A where the sum of the elements of B equals 2n;2 (2n + 1)
are allowed? Prove that one can nd x with exactly n questions
stated in advance.
Solution:
Lemma There exist n sets each with 2n;1 elements, the sum
of the elements of each set equals 2n;2 (2n + 1) with the following
property: the elements from the set f1 2 : : : 2n g get as answers
distinct n;tuples from "yes" and "no".
Proof: Induction by n 3. For n = 3 the sum of the elements
of B is 18 and we use the sets B1 = f1 2 7 8g B2 = f1 3 6 8g and
B3 = f1 4 6 7g: The table shows that the elements from A get as
answers distinct triples.
1 2 3 4 5 6 7 8
B1 = f1 2 7 8g + + - - - - + +
B2 = f1 3 6 8g + - + - - + - +
B3 = f1 4 6 7g + - - + - + + -
+ means "yes", ; means "no".
11
The number of elements in each set is 22 = 23;1 :
Suppose the assertion is true for some m: Therefore there exist
sets B1 B2 : : : Bm each with 2m;1 elements, the sum of the elements
of each set equals 2m;2 (2m +1) and every element from f1 2 : : : 2mg
gets distinct m;tuple.
Consider the set f1 2 : : : 2m+1g: For any i 1 i m if Bi =
fa1i a2i : : : a2m;1i g let
Di = fa1i a2i : : : a2m;1i a1i + 2m a2i + 2m : : : a2m;1i + 2m g:
It is clear that each set Di 1 i m has exactly 2m elements
and the sum of the elements of Di is equal to: 2(a1i + a2i + : : : +
a2m;1i)+2m;1 :2m = 2:2m;2 (2m +1)+22m;1 = 22m;1 +2m;1 +22m;1 =
2m;1 (2m+1 + 1): It is easily seen that only the elements t and t + 2m
for 1 t 2m get equal m;tuples. Let P and Q be nonintersecting
sets such that jP j = jQj = 2m;1 and P Q = f1 2 : : : 2m g: Consider
a set Q obtained from Q by adding 2m to each of its elements. The
set Dm+1 = P Q has 2m elements and the sum of its elements
is equal to s2P s + s2Q s + 2m;1 :2m = 2m;1 (2m + 1) + 22m;1 =
P P
2m;1 (2m+1 + 1): It is clear that exactly one element of each pair
(t t + 2m ) for 1 t 2m belongs to Dm+1 : Hence, we have found
the desired sets D1 D2 : : : Dm+1:
It is obvious now that the sets, given by the Lemma solve the
problem.
Problem 11.1. A sequence a1 a2 : : : an : : : is dened by
a1 = k a2 = 5k ; 2 and an+2 = 3an+1 ; 2an n 1
where k is a real number.
a) Find all values of k such that the sequence fang1n=1 is conver-
gent.
12
b) Prove that if k = 1 then
an+2 = 7a1n++1a; +8aan an+1 n 1
" #
2
n n+1
where x] denotes the integer part of x:
Solution: a) Write the given reccurent relation in the form an+2 ;
an+1 = 2(an+1 ; an) and consider the sequence cn = an+1 ; an:
It is obvious that if the sequence fang1n=1 is convergent then the
sequence fcng1 n=1 is also convergent. From the other hand it follows
from c1 = a2 ; a1 = 4k ; 2 and cn+1 = 2cn that the sequence fcn g1n=1
is an arithmetic progression and cn = (4k ; 2):2n;1 for any positive
integer n: Thus, if 4k ;2 6= 0 then the sequence fcn g1n=1 is unbounded
and therefore is not convergent one. Therefore 4k ; 2 = 0 i.e. k = 12 :
For this value of k all terms of the sequence fang1n=1 are equal to 21
and therefore the sequence is convergent.
b) It easily follows by induction that an = 2n ; 1: From the other
hand the equality from the condition of the problem is satised for
n = 1 2: Suppose it is true for n = m and n = m + 1: Then
7a2m+1 ; 8amam+1 = 7(2m+1 ; 1)2 ; 8(2m ; 1)(2m+1 ; 1) =
" # " #
1 + am + am+1 1 + 2m ; 1 + 2m+1 ; 1
3:22m+2 ; 2m+2 + ;1 = 2m+2 ; 1 = a
" #
m+1
3:2m ; 1 3:2m ; 1
which completes the proof.
Problem 11.2. On each side of a triangle with angles 30 60 and
90 and hypotenuse 1 is chosen a point such that the triangle formed
from these three points is a right triangle. What is the least value
of the hypotenuse of this triangle?
13
Solution: Consider triangle ABC with angles and . Let
A1 2 BC B1 2 CA and C1 2 AB be such that 4A1B1C1 is right
with <) A1C1B1 = 90 : Suppose that the circle k of diameter A1B1
intersects AB at point X 6= C1: It is easy to see that if 90
and 90 then X is an interior point for line segment AB: Draw
a tangent l to k which is parallel to AB _
and the tangential point
Y belongs to the smaller of the arcs C1X : Denote the intersecting
points of l and lines AC and BC by B2 and A2 respectively. Con-
sider homothety of center C for which the image of B2(A2) is the
point A(B ): It is obvious that the image of 4B1A1Y is inscribed
in 4ABC and its hypotenuse is less than B1A1: Therefore wlog we
may suppose that AB is tangent to k: In this case if <) B1A1C1 =
15
n
!
n+a;b
2
X n+a+b ! n;a;b !
The sum 2
i n;a+b ; i equals to the coecient in front
2
i=b 2
of x n;2a+b in the expansion of (1 + x) n+2a+b (1 + x) n;2a;b = (1 + x)n:
n
!
n n
! !
n+a+b n;a+b :
2 2
2 2
16
Spring mathematics tournament
Kazanlak, 30 March - 1 April 2001
Problem 8.1. Let a be a real parameter such that 0 a 1: Prove
that the solutions of the inequality
jxj + jax + 21 j 1
form an interval of length greater or equal to 1.
Solution: If a = 0 then all numbers from the interval 12 ; 12 of
4 4
Further, assume that 0 < a < 1 and let x 0: In this case
ax + 21 > 0 and therefore the solutions are x 2(1 1+ a) : Let x < 0:
If ax + 21 0 i.e. x ; 12 the solutions satisfy x ; 2(1 1; a) : If
ax + 21 < 0 i.e. x < ; 21a the solutions satisfy x ; 3(1 1+ a) :
Let a 12 : Then ; 2(1 1; a) ; 21a and so all numbers x
; 2(1 1; a) are solutions of the inequality. Since ; 3(1 1+ a) > ; 21a
it has no other solutions. Therefore all solutions form an interval
; 2(1 1; a) 2(1 1+ a) of length 2(1 1+ a) + 2(1 1; a) = 1 ;1 a2 > 1:
" #
Let a > 12 : Then ; 2(1 3+ a) < ; 21a and the solutions are x
17
; 2(1 3+ a) x ; 21a or x 2 ; 2(1 3+ a) ; 21a : Therefore all so-
" #!
18
b) Find the least integer n n > 1 for which
12 + 22 + + n2
n
is a perfect square.
Solution:
a) Direct verication shows that
n(n + 1)(2n + 1) + (n + 1)2 = (n + 1)(n + 2)(2n + 3)
6 6
for any integer n: Using this equality the proof is easily done by
induction.
b) It follows from a) that n(n + 1)(2n + 1) = 6m2: Since 2n + 1
is odd then n + 1 is even, i.e. n is odd. Let n = 2k ; 1: Then
k(4k ; 1) = 3m2: Therefore 3=k or 3=4k ; 1 i.e. 3=k or 3=k ; 1: Let
k = 3l: Then l(12l ; 1) = m2: Since (l 12l ; 1) = 1 we have l = n2
and 12l;1 = v2: The latter equality is impossible (contradiction both
modulo 4 and modulo 3). Let k = 3l +1: Then (3l +1)(4l +1) = m2:
Since (3l + 1 4l + 1) = 1 we get that 3l + 1 = u2 4l + 1 = v2 u >
1 v > 1: Verifying consequently for v = 3 5 7 9 11 13 : : : we obtain
l = 2 6 12 20 30 42 56 : : : : Thus, 3l + 1 = 7 19 37 61 91 127 169
(rst perfect square). So, the least n is 2k ; 1 where k = 3l + 1 =
3:56 + 1 = 169 and therefore n = 337:
Problem 9.1. Let f (x) = x2 + 6ax ; a where a is a real parameter.
a) Find all values of a for which the equation f (x) = 0 has at
least one real root.
b) If x1 and x2 are the real roots of f (x) = 0 (not necessarily
distinct) nd the least value of the expression
A = (1 +9xa ;)(14a+ x ) ; (1 ; 6a ;70xa)(1+;1 6a ; x ) :
2 3
1 2 1 2
19
Solution: a) The discriminant of f (x) is D = 4(9a2 + a). Therefore
the equation f (x) = 0 has at least one real root i 9a2 + a 0, so
giving a 2 ;1 ; 91 0 +1).
1 ; 7a
The quadratic function g(a) = 10a2 + 2a ; 1 attains its minimal
value for a = ; 10 1 62 ;1 ; 1 0 +1): Therefore the minimal
9
value of A equals to the smallest of the numbers g(0) and g ; 1 ,
1
89
i.e. this value is g ; 9 = ; 81 .
20
() OA:CO: QDOO = OB:OD () OA:CO = Q1O:OB:
1
Therefore the quadrilateral ABCQ1 is cyclic. Thus, Q1 Q:
Problem 9.3. Prove that there exist eight consecutive positive
integers such that non of them can be written in the form
j7x2 + 9xy ; 5y2j, where x and y are integers.
Solution: Denote f (x y) = 7x2 + 9xy ; 5y2. Since f (0 0) = 0,
f (0 1) = 5, f (1 0) = 7, f (1 1) = 11 and f (0 2) = 20, rst possible
sequence of eight positive integers is 12 13 : : : 19. We shall prove
that non of these integers can be written in the form j7x2 +9xy ;5y2j,
where x and y are integers.
Let f (x y) = k, where x and y are integers. It suces to prove
that f (x y) = k has no solutions for k 2 f12 13 : : : 19g.
Suppose k is even. Then x and y are also even. If x = 2x1 and
y = 2y1 we get the equality 4f (x1 y1) = k which implies that k is
divisible by 4. Thus, k 6= 14 and k 6= 18. Let k = 16 and consider
the equation 4f (x1 y1) = 16 which is equivalent to f (x1 y1) = 4.
As above we conclude that x1 and y1 are both even and let x1 = 2x2
and y1 = 2y2. Therefore f (x y) = 1. By analogy if k = 12 we get
the equation f (x y) = 3.
Multiply the equation f (x y) = k by 28 and write it in the
form
(14x + 9y)2 ; 221y2 = 28k:
Since 221 = 13:17 it is appropriate to consider modules 13 and 17.
Denote t = 14x + 9y and consider all possibilities for k, i.e.
k 2 f1 3 13 15 17 19g.
1) If k = 13 then t2 28:13 (mod 17) and so t2 7 (mod
17). Raising this congruence to 8-th power gives t16 (7)8 ;1
21
(mod 17) which is a contradiction to Fermat's theorem. Therefore
k 6= 13.
2) If k = 15 then t2 28:15 5 (mod 17). Raising in 8-th
power gives t16 (5)8 ;1 (mod 17), a contradiction.
The cases k = 17 19 1 3 are treated similarly.
Problem 10a.1. Let a and b be positive numbers such that both
of the equations (a + b ; x)2 = a ; b and (ab + 1 ; x)2 = ab ; 1 have
two distinct real roots. Prove that if the two bigger roots are equal
then the two smaller roots are also equal.
Solution: Let a and b satisfy the condition of the problem. Both
equations have two distinct roots i a > b and ab > 1: Since a > 0
it follows that a2 > ab > 1 i.e. a > 1: Therefore a > b > a1 and
a > 1: It follows from
p p of the problem thatpa + b +
p the condition
a ; b = ab + 1 + abp ; 1 i.e. pa ; b = (a ; 1)(b ; 1) + ab ; 1:
If a > b 1 then a ; b ab ; 1 () (a + 1)(b ; 1)
0 () b 1: Therefore b = 1: Conversely if 1 < b 1 then
p p a
a ; b ab ; 1 () (a + 1)(b ; 1) 0 () b 1 i.e. b = 1:
Thus, the two bigger roots are equal i b = 1 and a > 1: In this case
the two smaller roots are also equal.
Problem 10a.2. Let A1 and B1 be points respectively on the sides
BC and AC of 4ABC D = AA1 \ BB1 and E = A1B1 \ CD: Prove
that if <) A1EC = 90 and the points A B A1 E lie on a circle, then
AA1 = BA1:
Solution: Let F = AE \ BC: We prove that EA1 is the bisec-
tor of <) BEF which solves the problem. Indeed, then we have
<) BAA1 =<) BEA1 =<) FEA1 =<) ABA1 i.e. AA1 = BA1:
22
Using Ceva's and Menelaus's theorems for 4AA1C we obtain
AD A1F CB1 = 1 AD A1B CB1 = 1 and so
A1D CF AB1 A1D CB AB1
(1) A1F = CF
A1B CB
Let B 0 be a point on the ray A1B ! such that <) B 0EA1 =<) A1EF:
Then EA1 is the bisector of <) B 0EF and since <) A1EC = 90 it
follows that EC is the external bisector of the same angle. Therefore
A1F = CF and it follows from (1) that A1B = A1B 0 i.e. B = B 0
A1B 0 CB 0 CB CB 0
which completes the proof.
Problem 10a.3. Find all positive integers x and y such that
x3 + y3 ; x2y2
(x + y)2
is a nonnegative integer.
Solution: Let x and y be positive integers such that
z = x +y ;x y
3 3 2 2
(x + y)2
is a nonnegative integer. Substitute a = x + y and b = xy and write
the expression in the form b2 +3ab ; a2(a ; z) = 0: The discriminant
of this quadratic equation a2(4a + 9 ; 4z) is a perfect square, so
(4a + 9 ; 4z) = (2t + 1)2: Thus, a = t2 + t + z ; 2 and from the
equation for b we obtain that b = a(t ; 1): Since t 2 we have
(x ; y)2 = a2 ; 4a(t ; 1) < (a ; 2(t ; 1))2: From the other hand
a t2 and therefore a2 ; 4a(t ; 1) (a ; 2(t ; 1) ; 2)2: Since
a2 ; 4a(t ; 1) 6= (a ; 2(t ; 1) ; 1)2 (the two numbers are of dierent
parity) it follows that (x ; y)2 = a2 ; 4a(t ; 1) = (a ; 2(t ; 1) ; 2)2:
23
Thus, a = t2 i.e. t + z = 2 and so t = 2 z = 0 which implies
a = b = 4: Therefore x = y = 2 and these are the only positive
integers satisfying the condition of the problem.
Problem 10b.1. Solve the equation:
p
3log (cos x+sin x)+ ; 2log (cos x;sin x) = 2:
3
1
2 2
Solution: Note that the admissible values are those x for which
cos x + sin x > 0 cos x ; sin x > 0: After simple calculations the
equation becomes
p p p
( 3 ; 1) cos x + ( 3 + 1) sin x = 2:
This equation is equivalent to
p p p p
6 ; 2 cos x + 6 + 2 sin x = 1 :
4 4 2
p p p p
Since cos 15 = 6+4 2 and sin 15 = 6;4 2 we have
sin(x + 15 ) = 12 :
From the roots of the later equation only x = 15 k360 are ad-
missible. Therefore the roots are x = 15 k360 :
Problem 10b.2. Given a triangle ABC: Let M be such an interior
point of <) BAC that <) MAB =<) MCA and <) MAC =<) MBA:
Analogously, let N be such an interior point of <) ABC that
<) NBA =<) NCB and <) NBC =<) NAB and let P be such an inte-
rior point of <) ACB that <) PCA =<) PBC and <) PCB =<) PAC:
Prove that lines AM BN and CP intersect in a point on circumcircle
of 4MNP:
24
Solution: Suppose <) C < 90 : The case <) C > 90 is treated simi-
larly. Denote the intersecting point of line CP and line segment AB
by C1. Since PC1 is the bisector of <) APB (<) APC1 =<) BPC1 =
=<) ACB ) we get BCAC1 = AP : Triangles APC and BPC are sim-
1 BP
AP
ilar which gives CP = BC and BP
AC
CP AC= BC =) AP = AC 2 = b2 :
BP BC 2 a2
By analogy, if A1 and B1 are the intersecting points of AM and
BA c2 CB
BN with BC and AC respectively then A C = b2 and B A = c2 :
1 1 a 2
1 1
AC 1 :BA1:CA1
Thus, C B:A C:B A = 1 i.e. according to Ceva's theorem lines
1 1 1
AM BN and CP intersect in a point. Dnote this point by K: Since
<) APB =<) AOB = 2 where O is the circumcenter of 4ABC we _
get that A P O and B lie on a circle and PC1 intersects the arc AB
at its midpoint C1. Therefore OC2 is a diameter of this circle and
<) OPC2 =<) OPC = 90 : It easily follows now that M N and P lie
on a circle of diameter OK:
Problem 10b.3. Consider a set P of six four-letter words over an
alphabet of two letters a and b: Denote by QP the set of all words
over the same alphabet which do not contain as subwords the words
from P . Prove that:
a) if QP is nite then it does not contain words of length 11
b) there exists a set P such that QP is nite and it contains word
of length 10.
Solution: a) Suppose that QP contains a word of length 11. We
will show that QP contains words of any length. Let ! be a word of
length 11. There are 9 subwords of ! of length 3. Since there are 8
distinct words of length 3 it follows that there exists a word that
appears as subword of ! twise. Let = 123 where i 2 fa bg:
25
Consider the subword of ! obtained after the second appearance of
i.e. consider the word
: : : 1231 : : : 123
Write 1 after 3. Obviously the new word does not contain subwords
from P:
: : :2312 : : : 231
By analogy write 2 and so on. Thus, we nd words of any length
without subwords from P: This contradiction shows that there are
no words of length 11 in QP .
b) A direct verication shows that the set
P = f0000 1000 1001 1010 1101 1111 g
is such that QP is nite and 0001011100 2 QP :
Problem 11.1. Prove that there exist unique numbers and
such that cos = 2 tg = 1 and 0 < < < 1:
Solution: The function f (x) = cos x ; x2 is continuous in the in-
terval (0 1) f (0) = 1 > 0 and f (1) = cos 1 ; 1 < 0: Therefore there
exists such that cos = 2: In the interval (0 1) the function cos x
is decreasing one, whereas the function x2 is increasing. Therefore
there exists unique 2 (0 1) such that cos = 2:
We show now that in the interval 0 1] there exists a unique
such that g( ) = 0 where g(x) = xtgx ; 1. The function g(x) is
increasing one because xtgx is increasing as product of two increasing
functions. Moreover tg 4 = 4 ; 1 < 0 g(1) = tg1 ; 1 = tg1 ;
tg 4 > 0 and the uniquennes of follows. Since sin x < x for positive
26
x it follows that g() = tg ; 1 = sin ; 1 < 0: Further, g(x) is
increasing in the interval 0 1] and so < :
Problem 11.2. Let AA1 and BB1 be the altituds of obtuse non-
isosceles 4ABC and O and O1 are circumcenters of 4ABC and
4A1B1C respectively. A line through C intersects the line segments
AB and A1B1 at points D and D1 respectively and E is point on
the line OO1 such that <) ECD = 90 : Prove that EO 1
EO CD= CD1 :
(3) CO = 2 sin BC
<) BAC
(4)
CD1 = CA1 sin <) B1A1C = CA1 sin <) BAC
sin(<) B1A1C + <) D1 CA1) sin(<) BAC + <) DCB )
27
It follows now from (1),(2),(3),(4) and (5) that
EO1 = A1C sin <) BAC sin <) BDC = CD1
EO BC sin <) ABC sin(<) BAC + <) DCB ) CD
which completes the proof.
Problem 11.3. There are 2001 towns in a country every one of
which is connected with at least 1600 towns by direct bus line. Find
the largest n for which there exist n towns any two of which are
connected by direct bus line.
Solution: Let S1 snd S1 be two towns connected by direct bus line.
If k is the number of towns connected to both S1 and S2 by bus line
then (1599 ; k)+(1599 ; k)+ k 1999 which implies that k 1198:
Therefore there exists town S3 connected to both S1 and S2: Further,
let k be the number of towns connected to all S1 S2 and S3: Therefore
(1197 ; k) + (1598 ; k) + k 1998 so giving k 797: Therefore
there exists town S4 connected to all S1 S2 and S3: By analogy let k
be the number of towns connected to all S1 S2 S3 and S4: We have
(796 ; k) + (1597 ; k ) + k 1997 which implies k 396: Therefore
there exists S5 connected to all S1 S2 S3 and S4: For the number n
from the condition of the problem we obtain n 5: We show that
n = 5: For, number the towns by S1 S2 : : : S2001 and connect Sk
and Sm with direct line for all k and m for which k 6 m(mod5):
Since 2001
28
L National Mathematics Olympiad
3rd round, 28-29 April 2001
Problem 1. For which values of the real parameter a the equation
lg(4x2 ; (8a ; 1)x +5a2 )+ x2 +(1 ; 2a)x +2a2 = lg(x2 ; 2(a +1)x ; a2 )
has exactly one root?
Solution: Write the equation in the form:
30
Problem 3. Find the least positive integer n such that there exists
a group of n people such that:
1. There is no group of four every two of which are friends
2. For any choice of k 1 people among which there are no
friends there exists a group of three among the remaining n ; k
every two of which are friends.
Solution: Consider a group of 7 people A1 A2 : : : A7 such that
Ai i = 1 2 : : : 7 is not friend only with Ai+1 and Ai;1 (we set that
A8 = A1 and A0 = A7). It is easily seen that there are no four every
two of which are friends. Also, for any choice of k 1 people (in
this case k is 1 or 2) every two of which are not friends there exists
a group of three among remaining 7 ; k people every two of which
are friends. Therefore k 7:
We prove that for any group of 6 which satisfy the condition 1)
it is possible to choose a group of k 1 every two of which are not
friends such that among remaining 6 ; k a group of three every two
of which are friends does not exist.
Denote the people by A1 A2 : : :A6: If some of them is friend with
the other 5 (wlog suppose this is A1) it is clear that there are no three
among A2 A3 : : :A5 every two of which are friends. Therefore the
choice of A1 solves the problem.
Suppose one of then is friend with exactly four others (wlog as-
sume A1 is friend with A2 A3 A4 and A5) Then the choice of A1 and
A6 solves the problem.
Therefore each person has 0 1 2 or 3 friends. It is obviuos that
there exists a group of three any two of which are friends (otherwise
the problem is trivial). Assume that the group A1 A2 and A3 has
this property. Wlog A1 and A4 are not friends. Therefore among
A2 A3 A5 and A6 there is a group of three any two of which are
31
friends. If this is A5 and A6 together with one of A2 or A3 then there
exists a person who is friend with at least four others, a contradiction.
Therefore wlog suppose that this group is A2 A3 and A5. Since A1
and A5 are not friends and since among the others there is a group of
three friends we obtain that either A2 or A3 has at least four friends
which is a contradiction.
Therefore k = 7:
Problem 4. Given a right triangle ABC with hypotenuse AB:
A point D distinct from A and C is chosen on the ray AC ! such
that the line through incenter of 4ABC parallel to the bisector of
<) ADB is tangent to the incircle of 4BCD: Prove that AD = BD:
Solution: First we show that C lies on the line segment AD: For,
suppose the contrary, i.e. <) ADB > 90 : Denote the tangential
point of incircle k(I r) of 4BCD and the side BD by P and the
tangential point of k and the line through the center J parallel to
bisector l of <) ADB by T . Since l ? DI we get T 2 DI: Thus,
<) IJT = <) CBD2 =<) IBP and since IT = r = IP we have that
4IJT and 4IBP are congruent. In particular IJ = IB which
implies that <) BJC < 90 a contradiction. Let A0 be such point on
the ray DA! that DA0 = DB and E is the midpoint of A0B: Denote
the tangential point of incircle of 4A0BC by F and its incenter by
J 0. Then we have
EF = jA0E ; A0F j = 21 jA0B ; (A0B + A0C ; BC )j =
= 12 jBC ; A0C j = 12 jBC ; (A0D ; CD)j =
1
2
jBC + CD ; BDj = r:
Therefore the line J 0F parallel to l is tangent to k: This implies
that J 0 2 JF: From the other hand J 0 2 JC and since JC is not
32
parallel to JF (<) ACJ = 45 > 21 <) ADB =<) ADI ) we get J = J 0:
Thus, <) ABJ =<) CBJ =<) A0BJ i.e. A = A0 which completes the
proof.
Problem 5. Find all triples of positive integers (a b c) such that
a3 + b3 + c3 is divisible by a2b b2c and c2a:
Solution: If d = gcd(a b c) it is easy to see that da db dc is
!
33
problem.
Supose now that c > b 3: Since 1 + b3 + c3 b2c2 it follows
that 2c3 > 1+ b3 + c3 and so 2c > b2 or c > b2 : It follows from 2c > b2
2
that 2c > b2 ; b + 1 i.e. (3) b ;cb + 1 < 2: From the other hand
2
Multiplying (3) and (4) gives b c+2 1 < 1 which is impossible since
3
n n
n 0.
Solution: Let xn = an +2an;1 , x0 = 3 and yn = an ;2an;1 , y0 = 1.
Then xn = 3xn;1 +4yn;1 and yn = 2xn;1 +3yn;1 . Since an = xn +yn,
it suces to prove that xn + yn = xyn ;+y7 , i.e. x2n = 2yn2 + 7.
2
n n
We prove this by induction. The assertion is obvious for n = 0.
Suppose that x2n;1 = 2yn2;1 + 7. Writing this equality in the form
(3xn;1 + 4yn;1)2 = 2(3yn;1 + 2xn;1)2 + 7 gives x2n = 2yn2 + 7 which
completes the proof.
Problem 2. Given nonisosceles triangle ABC . Denote the tan-
gential points of the inscribed circle k of center O with the sides
AB BC and CA by C1 A1 and B1 respectively. Let AA1 \ k =
A2 BB1 \ k = B2 and let A1A3, B1B3 be bisectors in triangle
A1B1C1 (A3 2 B1C1 B3 2 A1C1). Prove that:
a) A2A3 is bisector of <) B1A2C1
b) if P and Q are the intersecting points of circumcircles of tri-
angle A1A2A3 and triangle B1B2B3 then the point O lies on the line
PQ.
Solution: a) From the Low of Sine's we obtain
AB1 = sin <) B1A1A2 AC1 = sin <) C1A1A2
AA1 sin 1 AA1 sin 1
35
where 1 =<) A1B1C1 1 = A1C1B1: Since AB1 = AC1 we get
A2B1 = sin <) B1A1A2 = sin 1 = A1B1 = A3B1
A2C1 sin <) C1A1A2 sin 1 A1C1 A3C1
which implies that A2A3 is the bisector of <) B1A2C1:
b) Let M = AA1\BB1: It follows from MA1 : MA2 = MB1 : MB2
that M lies on the line PQ. Therefore it suces to prove that
OM ? O1O2, where O1 and O2 are circumcenters of 4A1A2A3 and
4B1B2B3. It follows from a) that the diametrically opposite point
of A3 in k1 { the circumcircle of 4A1A2A3, lies on the line B1C1.
Therefore O1 2 B1C1. Moreover <) B1B3A1 =<) CA1A3 = + 1 .
2
It easily follows now that O1 coinsides with the intersecting point
of B1C1 and BC . Let OO1 \ A1A2 = N and OO2 \ B1B2 = K . It
follows from 4OA1O1 that ON : OO1 = OA21 = r2 and by analogy
OK : OO2 = r2 where r is the radius of k. Since O, N , M and K lie
on the circle k3 of diameter OM we have that the line O1O2 is the
image of k3 by inversion of center O and degree r2, i.e. OO1 ? OM .
Problem 3. For a permutation a1 a2 : : : an of the numbers 1 2 : : : n
it is allowed to change the places of any two consecutive blocks, i.e.
from
a1 : : : ai ai+1 ai+2 : : : ai+p ai+p+1 ai+p+2 : : : ai+q ai+q+1 : : : an
| {z } | {z }
A B
by replacing A and B one can obtain
a1 : : : ai ai+p+1 ai+p+2 : : : ai+q ai+1 ai+2 : : : ai+p ai+q+1 : : : an.
| {z } | {z }
B A
Find the least number of such changes after which from n n ;
1 : : : 1 one can obtain 1 2 : : : n.
Solution: Call the change of two blocks a move. We shall prove
that the least number of moves such that from n n ; 1 : : : 1 one can
36
obtain 1 2 : : : n is n +2 1 .
Consider the number of pairs ai ai+1 such that ai < ai+1. This
number is 0 in the initial permutation n n ; 1 : : : 2 1 and is n ; 1
in the nal permutation 1 2 : : : n ; 1 n.
First we show that a move changes the number of pairs ai ai+1
such that ai < ai+1 at most by two. For, consider a move of the
blocks a : : : b and c : : : d of the permutation
: : : p a : : : b c : : :d q : : :
As a result we obtain
: : : p c : : : d a : : :b q : : :
It is clear that at most three pairs can change the ordering. Suppose
the elements in all three pairs change the ordering, i.e. from p > a,
b > c and d > q we get p < c, d < a and b < q. Adding the
rst three inequalities gives p + b + d > a + c + q and adding the
last three implies p + b + d < a + c + q, a contradiction. Therefore
a move changes the number of pairs ai ai+1 for which ai < ai+1
at most by two. It is easily seen that the rst and the last moves
change this number by one. Therefore if x is desired number we have
2 + 2(x ; 2) n ; 1 which implies x n +2 1 . It remaines to nd
39
have
2 2 p
p1 + p2 ; ( t)2
(1) 3
2 p
3
1 p2 = cos <) ACB > 21
3 3
p 2 2
p
( t)2 + p13 ; p23
(2) 2p p2 = cos <) BAC > ; 23 :
3
p
These inequalities
p are equivalent to t < 1 and t + t ; 1 > 0 i.e.
3 ; 5
!
t2 2 1 :
p
Conversely, let t 2 3 ;2 5 1 : Then we can construct a tri-
!
p
angle ABC such that AB = t BC = p2 CA = p1 : By (1) and
3 3
(2) it follows that <) ACB < 60 and <) ABC <<) BAC < 150 : Let
kA be the circle through B and C such that the arc BC , lying in
the half-plane (with respect to BC ) containing the point A is equal
to 60 : Denote by kB the analogous circle trough A and C: Then it
is easy to see that the second common point O of kA and kB lies
in the interior of 4ABC: (Indeed, assume the contrary. If O lies in
<) ACB then <) AOB =<) AOC + <) BOC = 150 + 150 = 300
which is impossible. If O lies in the opposite angle of <) ACB
then <) AOB =<) AOC + <) BOC = 30 + 30 = 60 which con-
tardicts to the inequalities <) AOB <<) ACB < 60: The other cases
for the position of O can be rejected in the same manner.) Hence
<) AOB = 360 ; <) AOC ; <)pBOC = 360 ; 150 ; 150 = 60:
Set x = AO y = BO z = CO 3: Applying the cosine theorem for
triangles AOB BOC and COA it follows that x y z t satisfy the
conditions of the problem.
40
Problem 6. Given the equation
(p + 2)x2 ; (p + 1)y2 + px + (p + 2)y = 1
where p is xed prime number of the form 4k + 3: Prove that:
a) If (x0 y0) is a solution of the equation where x0 and y0 are positive
integers, then p divides x0
b) The given equation has innitely many solutions (x0 y0) where
x0 and y0 are positive integers.
Solution: a) Set y ; 1 = z and write the equation in the form
(3) x2 = (z ; x)((p + 1)(z + x) + p):
If z ; x and (p + 1)(z + x) + p are relatively prime then they are
perfect squares which is impossible since the second one is of the
form 4k + 3: Let q be a common divisor of these numbers. It follows
from (3) that q=x and so q=z: Since q=(p + 1)(z + x) + p we have q=p
giving q = p which completes the proof.
b) It suces to prove that (3) has innitely many solutions in
positive integers. Let x = px1 and z = pz1: Then x21 = (z1 ; x1)((p +
1)(z1 + x1) + 1) and therefore there exist positive integers a and b
such that z1 ; x1 = a2 x1 = ab and (p + 1)(z1 + x1) + 1 = b2: It
follows now that
(4) (p + 2)b2 ; (p + 1)(a + b)2 = 1:
q q 2k+1 q q
Let p + 2 + p + 1 = mk p + 2 + nk p + 1 for any k =
0 1 : : : where mk and nk are positive integers. It is obvious that
q q q q
( p + 2 ; p + 1)2k+1 = mk p + 2 ; nk p + 1
and after multiplying we obtain that (p+2)m2k ;(p+1)n2k = 1 i.e. b =
mk and a + b = nk are solutions of (4): Hence, x = pmk (nk ; mk) and
41
z = pnk (nk ;mk ) are solutions of (3): The assertion of b) follows from
the fact that both sequences m1 m2 : : : and n1 n2 : : : are strictly
increasing.
42
SPRING MATHEMATICAL COMPETITION
1995
Grade 8.
Problem 1. Find all values of a, for which the system
x + 4jyj = jxj
jyj + jx ; aj = 1
has exactly two solutions.
Solution. Let (x y ) be a solution with x 0. Then y = 0 and jx ; aj = 1, i.e. x = a 1. It
is obvious that when a 1 the system has two solutions with x 0, namely (a ; 1 0), (a +1 0).
When ;1 a < 1 the system has only one solution (a + 1 0) with x 0. Let x < 0. Then
jyj = ; x2 and consequently jx ; aj = 1 + x2 . Since jx ; aj 0, then 1 + x2 0 or x ;2.
We have x ; a = 1 + x2 and x = 2(a + 1), x ; a = ;1 ; x2 . From here x = 23 (a ; 1). From
;2 2(a + 1) < 0 we get;2 a < ;1 and from ;2 23 (a ; 1) < 0 we get ;2 a < 1.
Obviously if (x y ) is a solution of the system with x < 0, then y 6= 0 and therefore (x ;y )
is a solution too. Thus, when a < ;2 the system has no solution. When a = ;2 we get
x = 2(a + 1) = ;2 and x = 32 (a ; 1) = ;2, i.e. the system has exactly two solutions (;2 1).
If 1 > a > ;2, then 2(a + 1) 6= 32 (a ; 1) and the two values of x give solutions, thus we get four
solutions with x < 0.
Therefore the system has two solutions only when a 1, namely (a ; 1 0) and (a + 1 0) and
when a = ;2, namely (;2 1).
Problem 2. Let M be the midpoint of the side BC of the parallelogram ABCD, N be the
common point of AM and BD, while P be the common point of AD and CN . Prove that
a) AP = AD
b) CP = BD i AB = AC .
Solution. a) Consider 4ABC . The lines AM and BD are medians and thus N is the center
of gravity. If Q is the intersection point of CP and AB , then Q is the midpoint of AB . It's easy
to see that 4APQ = 4BCQ (Figure. 1), from where AP = BC = AD.
b) Let AB = AC . Then 4ABC is isosceles and AM is a median in it. It is easy to
see that 4NBC is isosceles and BN = CN . Analogously NA is a median and an altitude
in 4PND, thus PN = DN , i.e. PC = PN + CN = DN + BN = BD (Figure 2). Let
CP = BD. Trough B we draw a line BF , parallel to CP . Obviously PFBC is a parallelogram
and therefore 6 CPD = 6 BFP . Since CP = BF , then 4DBF is isosceles. Consequently
1
6 NDA = 6 BFP = 6 CPD and DN = NP . A is the midpoint of PD and NA is perpendicular
to AD and BC . We get that AM is perpendicular to BC and we deduce from here that 4ABC
is isosceles, i.e. AB = AC .
Figure 1. Figure 2.
Let now M be the set of all pairs, that are consisted of a side and a vertex from which this
side is seen under an acute angle. Let (AB , C ) be an element of M . Consider the circumcircle of
4ABC and let N be the set of all such circles. We denote by k the circle in N with the biggest
radius. Let k be the circumcircle of 4A1A2 Am , where A1 A2 is a side of the polygon (Figure
2
4). We shall show that k is the circle we are looking for. Assume that there exists a vertex Ap
which is outside k. Then 6 A1 Ap A2 < 6 A1Am A2 < 90 and the circumcircle of 4A1A2 Ap is
from N and its radius is greater than the radius of k (Lemma 1). We get a contradiction.
Consider one of the acute angles in 4A1A2 Am . Let = 6 A2 A1 Am be acute. We shall prove
that Am = A3 , which means that k passes through 3 consecutive vertices of the polygon. Assume
that Am 6= A3. Then A3 is situated in the way which is shown in the Figure 4. If = 6 A2 A3 Am ,
then + > 180, i.e. 180 ; < < 90. Thus 6 A2 Am A3 < 90, which means that (A2A3 ,
Am ) is an element of M . Let c be the circumcircle of 4A2 A3Am and X be the intersection point
of c and the segment bisector of A2 Am . Since 6 A2 XAm = 180 ; < = 6 A2 A1Am , then the
radius of c is greater than the radius of k (Lemma 1). This is a contradiction.
Grade 9.
Problem 1. Let M be an arbitrary point on the side AB = 1 of the equilateral triangle
ABC . The points P and Q are orthogonal projections of M on AC and BC , while P1 and Q1
are orthogonal projections of P and Q on AB .
a) Prove that P1 Q1 = 43 .
b) Find the position of M for which the segment PQ is with the smallest length.
Solution. a) We have SABC = SACM + SBCM = Figure 5.
1 (AC:MP + BC:MQ) = 1 (MP + MQ) (Figure 5). On
2 p2 p
the other hand SABC = 4 . Thus MP + MQ = 23 .
3
p triangles P1MPp and MQ1Q
Now from the rectangular
we evaluate P1 M = 23 MP and MQ1 = 23 MQ. From
p
here P1 Q1 = P1 M + MQ1 = 23 (MP + MQ) = 43 .
b) The orthogonal projection of the segment PQ on
AB is the segment P1Q1 and thus PQ P1 Q1 . There-
fore PQ is with minimal length when it is parallel to
P1Q1. The last is true exactly when AP = BQ. We get
that 4AMP = 4BMQ and hence AM = BM . So, PQ
is minimal when M is the midpoint of AB .
Problem 2. The quadratic function f (x) = ;x2 + 4px ; p + 1 is given. Let S be the area
of the triangle with vertices at the intersection points of the parabola y = f (x) with the x-axis
and the vertex of the same parabola. Find all rational p, for which S is an integer.
Solution. The discriminant of f (x) is D = 4(4p2;p+1) and D > 0 for all real p. Consequently
f (x) has two real roots x1 and x2, i.e. f (x) intersects the x-axis in two dierent points | A
and B . The vertex C of the parabola has coordinates 2p and h = f (2p) = 4p2 ; p + 1 > 0. We
have q q q
AB = jx1 ; x2j = (x1 ; x2 )2 = (x1 + x2 )2 ; 4x1x2 = 2 4p2 ; p + 1:
Now we nd S = S = AB:h = (4p2 ; p +1) 2 : Denote q = 4p2 ; p +1. Since q is rational
3
ABC 2
and q 3 = S 2 is an integer, then q is an integer too. Then S is rational and S 2 = q is integer,
q q
3
thus Sq is integer too. Therefore q = n2 , where n is a positive integer, i.e. 4p2 ; p + 1 ; n2 = 0.
The quadratic equation (with respect to p) has a rational root exactly when its discriminant
16n2 ; 15 is a square of a rational number. Consequently 16n2 ; 15 = m2 , and we can consider
m to be a positive integer. From the equality (4n ; m)(4n + m) = 15 we get 4n ; m = 1,
4n + m = 15 or 4n ; m = 3, 4n + m = 5. From here n = 2, m = 7 or n = 1, m = 1. The rational
numbers we are looking for are 0 1 41 ; 34 .
Problem 3. Let n be a positive integer and X be a set with n elements. Prove that
a) The number of all subsets of X (X and included) is equal to 2n .
b) There exist 2n;1 subsets of X each pair of which is with common element.
c) There do not exist 2n;1 + 1 subsets of X , each pair of which is with common element.
Solution. a) We use induction with respect to n. The base of the induction is obvious.
Assume that the assertion is true for a set with n ; 1 elements and let X be with n elements.
We can assume that X = f1 2 : : : ng. Let Y = f1 2 : : : n ; 1g. All subsets of X are divided
into two groups: I group | those which do not contain n and II group | those, which contain
n. Both groups have one and the same number of elements because each set of the II group is
obtained from exactly one set of the I group by annexing n. Thus the number of the elements
of X is twice greater than the number of the subsets of the I group. But the subsets of the I
group are exactly the subsets of Y and according to the inductive assumption their number is
2n;1 . Thus the number of the subsets of X is 2n .
b) According to a) the number of the subsets of X from the II group is 2n;1 and each pair
of them has a common element | the number n.
c) If A
X , let A = X n A. All subsets of X are divided into pairs fA Ag and the number of
these pairs is 2n;1 . Now if we have 2n;1 + 1 arbitrary subsets of X , according to the pigeonhole
principle, it is not possible that they are in dierent pairs of the type fA Ag. Consequently
among the given 2n + 1 subsets there exist two pairs of the type fA Ag, which obviously have
no common element.
Grade 10.
Problem 1. Find all values of the real parameters p and q, for which the roots of the
equations x2 ; px ; 1 = 0 and x2 ; qx ; 1 = 0 form (in a suitable order) an arithmetic
progression with four members.
Solution. Denote by x1 , x2 the roots of the equation x2 ; px ; 1 = 0 and by y1 , y2 the
roots of x2 ; qx ; 1 = 0. It is clear that for all p and q the numbers x1 , x2, y1 , y2 are real
and x1 x2 = y1 y2 = ;1. Assume that x1 < 0 < x2 and y1 < 0 < y2 . If four numbers a, b, c
and d form an arithmetic progression then the numbers d c b a form an arithmetic progression
too. So, we can assume that x1 , x2, y1 , y2 in a suitable order form an increasing arithmetic
progression. If x1 < y1 (the case y1 < x1 is analogous) then there are two possibilities:
I. The arithmetic progression is x1, y1, y2, x2. Then x1 + x2 = y1 + y2, from where p = q,
i.e. x1 = y1 , x2 = y2 which is impossible.
II. Theparithmetic progression is x1, y1, x2, y2. Then x2 ; x1 = y2 ; y1, from where
p2 + 4 = q2 + 4, i.e. p2 = q 2 and since p 6= q, then p = ;q 6= 0.
p
2
In the same way we have y1 = x1 +2 x2 = p2 and therefore p4 ; pq 2 ; 1 = 0. Since p = ;q ,
then p2 = 43 and p = p2 . Hence (p q ) = p2 ; p2 , ; p2 p2 .
3 3 3 3 3
4
Problem 2. Triangle ABC with AB = 22, BC = 19, CA = 13 is given.
a) If M is the center of gravity of 4ABC , prove that AM 2 + CM 2 = BM 2 .
b) Find the locus of points P from the plane of 4ABC , for which AP 2 + CP 2 = BP 2 .
c) Find the minimal and maximal values of BP , if AP 2 + CP 2 = BP 2 .
Solution. a) We have Figure 6.
p
AM = 23 ma = 23 21 2b2 + 2c2 ; a2 = 105
p
CM = 23 mc = 23 21 2a2 + 2b2 ; c2 = 8
p
5
last note that the product of all numbers (
) is equal to 38 :54:74:194. The only possibility is
a1a2 : : :a7 = 36 :54:74:194. From here a2 a3 : : :a7 = 36:53:73:193 and the only possibility is
fa2 a3 : : : a7g = f3 5 3 7 3 19 3 5 7 3 5 19 3 7 19g:
Therefore n = 7 and
fa1 a2 a3 a4 a5 a6 a7g = f3 5 3 7 3 19 3 5 7 3 5 19 3 7 19 5 7 19g:
Grade 11.
!
1 2 n
Problem 1. Let an = n2n++11 21 + 22 + + 2n , n = 1 2 3 : : :. Prove that
a) an+1 an for all n 3
b) the sequence fan g1n=1 is convergent and nd its limit.
Solution. a) We have
!
1 2 n+1
an+1 = n2n++22 21 + 22 + + n2 + 1 = 2(nn++21) (an + 1):
From here
2 ; an) ; (an + 1) n = 1 2 3 : : ::
an+2 ; an+1 = (n + 2) 2((ann+1
+ 1)(n + 2)
Since an > 0 for all n, then if an+1 ; an 0, we have an+2 ; an+1 0. But a3 = 35 and
a4 = 53 , i.e. a4 ; a3 = 0. Hence a5 ; a4 0, a6 ; a5 0, : : :, an+1 ; an 0.
Figure 7. b) The sequence a1 a2 a3 : : : is decreasing when n
3 and it is bounded (an > 0 for all n). Therefore this
sequence is convergent. Let nlim !1 n
a = a. From the
n + 2
equality an+1 = 2(n + 1) (an +1) after passing to innity
we get a = 12 (a + 1), i.e. a = 1.
Problem 2. The quadrilateral ABCD is inscribed
in a circle with center O. The diagonals AC and BD
intersect each other in the point M , M 6= O. The line
through M which is perpendicular to OM intersects the
sides AB and CD of the quadrilateral ABCD in the
points X and Y , respectively. Prove that AB = CD i
BX = CY .
Solution. If AB = CD, then ABCD is isosceles
trapezoid. Hence OM ?AD and OM ?BC , from where
XY kBC and BX = CY .
Let BX = CY . Denote by P and Q the midpoints of AB and CD, respectively. The quadri-
laterals OPXM and OQY M are inscribed, thus 6 OPM = 6 OXM and 6 OQM = 6 OY M .
We shall prove that 6 OPM = 6 OQM . The triangles ABM and DCM are congruent and
6
MP and MQ are medians in them. Therefore 4MPB 4MQC and 6 MPB = 6 MQC .
Then 6 OPM = 90 ; 6 MPB = 90 ; 6 MQC = 6 OQM: (Or 6 OPM = 6 MPB ; 90 =
6 MQC ; 90 = 6 OQM .)
Hence 6 OXM = 6 OY M and OX = OY . Therefore 4OXB = 4OY C . From here 6 OBA =
6 OCD and the isosceles triangles ABO and DCO are equal, i.e. AB = CD
8
Problem 2. The sides AC and BC of the triangle ABC are diameters of two circles, each
of which touches internally a circle k, which is concentric to the incircle of 4ABC .
a) Prove that AC = BC .
b) If cos 6 BAC = 34 , nd the ratio of the radii of k and the incircle of 4ABC .
Solution. a) Let I be the center of the incircle of 4ABC , N be the common point of
this circle with AC and M be the midpoint of AC . Let r1 be the radius of k and r be the
radius of the incircle of 4ABC . From the condition it follows that IM = 2b ; r1 and from the
2
MN = jAM ; AN j = jc ;2 aj . Therefore
(c ; a)2 = (b ; 2r1)2 ; 4r2 : (1)
Analogously
(c ; b)2 = (a ; 2r1)2 ; 4r2 : (2)
Assume that a 6= b. From (1) and (2) we get (b ; a)(2c ; a ; b) = (b ; a)(a + b ; 4r1), i.e.
2c ; a ; b = a + b ; 4r1. From here b ; 2r1 = c ; a and from (1) it follows that r = 0, which is
impossible. Thus a = b.
b) Let rr1 = t and 6 BAC = . Then c = 2b cos = 23 b and
s
Since b > 2r1, from (1) it follows that t = b ; (c ;2rb) + 4r and substituting c and r we
p
2 2
p
get t = 23 ( 7 ; 2).
Problem 3. n points (n > 4), no three of which are colinear are given in the plane. More
than n triangles are constructed with vertices among these points. Prove that at least two
triangles have exactly one common vertex.
Solution. Assume the contrary and let k be the smallest number for which the assertion is
not true. This means that there are constructed at least (k + 1) triangles using k points. It
follows from the pigeonhole principle that there exists a point A which is a vertex of at least 4
triangles. Let ABC be the rst triangle. At least one of the points B and C is a vertex of the
second triangle, which we denote by ABD. If ACX is the third triangle then X = D. Thus the
forth triangle must contain B or C , which is impossible. Therefore if A and B are vertices of two
triangles then they are vertices of all the four triangles. Let A be a vertex of t triangles, t 4.
These triangles are of the kind ABA1 , ABA2 , : : :, ABAt , where all the points A1 A2 : : : At
are pairwise dierent. Obviously it is not possible to exist a triangle of the type BXY , where
X and Y are points which are dierent from A1 A2 : : : At. Triangles BAi Aj and Ai Aj Am do
not exist too. Hence the points A, B , A1 , A2 , : : :, At are vertices only of the triangles ABA1 ,
ABA2 , : : :, ABAt . In such a way we use t + 2 points and get t triangles. It is not possible that
t + 2 = k, because all triangles are t < k. The number of remaining points is k0 = k ; t ; 2 and
by them there are constructed at least k + 1 ; t > k0 triangles, such that no two of them have
9
exactly one common vertex. The triangles are more than the points k0, and thus k0 > 4. We
have found a number k0 < k for which the assertion is not true. This contradicts the choice of
k.
10
SPRING MATHEMATICAL COMPETITION
1996
Grade 8
Problem 1. Prove that for all real a 2 (1 2) the area of the gure encountered by the
graphs of the functions y = 1 ; jx ; 1j and y = j2x ; aj is less than 31 .
Solution. Firstly, we shall nd the common points of the given functions. For this purpose
we solve the equation
j2x ; aj = 1 ; j1 ; xj: (1)
a a a
Since 1 < a < 2, then 2 < 1, and we shall consider the cases: x 2 , 2 < x < 1 and x 1.
We have:
1. When x a2 the equation (1) takes the form a ; 2x = x. Then x = a3 , which satises
(1), because a3 < a2 .
2. When a2 < x < 1 the equation (1) takes the form 2x ; a = x, i.e. x = a, which does not
satisfy (1), because a > 1.
3. When x 1 the equation (1) takes the form 2x ; a = 2 ; x. Then x = a +3 2 , which
satises (1), because a +3 2 > 1 when a > 1.
Thus, the graphs of the two functions have two common Figure 1.
points (Figure 1), the rst of which (denoted by A) has coor-
dinates xA = a3 , yA = a3 , while the second one (denoted by B )
has coordinates xB = a +3 2 , yB = 2 ; a +3 2 = 4 ;3 a . Denote
by C , D and E the points, with coordinates xC = 1, yC = 1
xD = a2 , yD = 0 and xE = 2, yE = 0, respectively.
Then, the gure encountered by the two graphs is the
quadrilateral ACBD and its area S is obtained by subtracting
the areas of the triangles ODA and BDE from the area of the
triangle OEC . Therefore,
S = SOEC ; SODA ; SBDE
= 1 ; 12 OD yA ; 12
1 DE y
B
= 1; 1 a a ; 1 2; a 4;a
2 2 3 2 2 3
1
= 1 ; 12 1 a2 ; 1 (4 ; a)2
2 12
1
= ;a + 4a ; 2
6
= 16 2 ; (a ; 2)2 < 13 :
Figure 2. Problem 2. The altitude CD of the rectangle triangle
ABC (6 ACB = 90) is a diameter of the circle k, which meets
the sides AC and BC in E and F , respectively. The intersec-
tion point of the line BE and the circle k, which is dierent
from E , is denoted by M . Let the intersection point of the lines
AC and MF be K , and the intersection point of the lines EF
and BK be P .
a) Prove that the points B , F , M and P are concyclic
b) If the points D, M and P are colinear, nd the angles
A and B of the triangle ABC .
Solution. a) Since 6 ECF = 90, then EF is a diameter of
the circle k and consequently 6 EMF = 90 = 6 BMK . Then
BC and KM are altitudes in 4BEK (Figure 2), which means
that the point F is the altitude center of this triangle. It follows from here that EP ? BK , i.e.
6 BPF = 90.
Thus, 6 BPF = 6 BMF = 90 and consequently the points B , F , M and P lie on a circle
with diameter BF .
Figure 3. b) We have:
6 BDM = 90 ; 6 MDC = 90 ; 6 MEC
Therefore 6 AXB is the biggest angle in 4XAB , from where 6 AXB > 60. Now, if we
assume that the town X is connected with at least 6 towns, then the sum of the angles at X
would be greater than 6 60 = 360, which is impossible. Thus, every town is connected with
at most 5 other towns.
c) Assume that there exists a closed piecewise line A1A2 : : :An , consisted of ways (Figure
6). The distances A1 An and A1 A2 are dierent. Let A1An < A1A2 . Then A2 is not the nearest
town to A1 and consequently (because the way A1 A2 exists) A1 is the nearest town to A2 . It
follows from here that A1 A2 < A2 A3 .
Proceeding in this way we obtain the chain: A1 An < A1A2 < A2 A3 < : : : < An A1 , which
leads to contradiction.
Grade 9
Problem 1. Find the values of the real parameter b, for which the dierence between the
maximal and the minimal values of the function f (x) = x2 ; 2bx +1 in the interval 0 1] is equal
to 4.
Figure 7. Figure 8. Figure 9.
- - -
b 0 1 0 1 b 0 b 1
Solution. It is clear that the minimal value of the quadratic function f (x) is obtained when
x = b. We shall consider the following three cases:
1. Let b < 0. In this case the function f (x) is increasing in the interval 0 1] (Figure 7)
and the maximal value is f (1) = 2 ; 2b, while the smallest one is f (0) = 1. From the condition
f (1) ; f (0) = 1 ; 2b = 4 we nd b = ; 23 , which is a solution of the problem.
3
2. Let b > 1. Now the function f (x) is decreasing in the interval 0 1] (Figure 8) and
the maximal value is f (0) = 1, while the minimal one is f (1) = 2 ; 2b. From the condition
f (0) ; f (1) = 2b ; 1 = 4 we nd b = 25 , which is a solution of the problem.
3. Let 0 b 1 (Figure 9). The minimal value of f (x) in the interval 0 1] is f (b) = 1 ; b2,
while the the maximal one is f (0) or f (1). From f (0) ; f (b) = b2 = 4 we nd b = 2, which are
not solutions (because in this case 0 b 1). From f (1) ; f (b) = (1 ; b)2 = 4 we nd b = 3 or
b = ;1, which are not solutions too.
Finally, the answers are: b = ; 23 and b = 52 .
Figure 10. Problem 2. The quadrilateral ABCD is inscribed in a
circle with radius 1, a circle can be inscribed in it and AB =
AD. Prove that:
a) the area of the quadrilateral ABCD does not exceed 2
p b) the inradius of the quadrilateral ABCD does not exceed
2.
2
Solution. We have (Figure 10) AB + CD = AD + BC and
from AB = AD we get: BC = CD. Let AB = AD = x, BC =
CD = y. The triangles ACB and ACD are equal (by SSS),
from where 6 B = 6 D. But ABCD is inscribed in a circle and
consequently 6 B + 6 D = 180. Therefore 6 B = 6 D = 90.
Then AC is diameter of the circumcircle of the quadrilateral and particularly AC = 2 and
x2 + y2 = AC 2 = 4.
Let S , p and r be the area of the quadrilateral, its semiperimeter and inradius, respectively.
a) We have S = SABC + SACD = 2SACB = xy and
q
xy = x2y2 x +2 y = AC2 = 2
2 2 2
Now
p
r 22 () r2 12 () 4 +S 2S 12
2
() S 2 ; S ; 2 0 () (S + 1)(S ; 2) 0
() S 2
p S > 0). The last inequality is true according to a) and consequently the inequality
(because
r 22 is true too.
Problem 3. This problem is the same as problem 3, grade 8.
4
Grade 10
Problem 1. Find in the plane the locus of points with coordinates (x y), for which there
exists exactly one real number z , satisfying the equality:
xz4 + yz3 ; 2 (x + jyj) z2 + yz + x = 0:
Solution. Let (x y ) be the coordinates of a point from the locus, we are looking for. This
means that there exists exactly one real number z , for which
xz4 + yz3 ; 2 (x + jyj) z2 + yz + x = 0:
If x = 0, the above equality takes the form
z yz2 ; 2jyjz + y = 0:
This equality is satised for at least two dierent values of z (for example z = 0 and z = 1,
if y 0 and z = 0 and z = ;1, if y < 0), which shows that the condition of the problem is not
veried. Consequently, the points from the y -axis do not belong to the locus.
Let x 6= 0. Thus, if z satises the given equality, then z 6= 0. It is easy to see that in this
case the number z1 satises the same equality and consequently z = z1 . From here z 2 = 1, i.e.
z = 1.
Case 1. Let z = 1. Then x + y ; 2 (x + jyj) + y + x = 0, from where; y ; jyj = 0. This
shows that y 0. If y = 0 and then the given equality takes the form x z 2 ; 1 2 = 0, and
consequently it is satised also by z = ;1. Thus, we can assume that y > 0. Then,
xz4 + yz3 ; 2(x + y)z2 + yz + x = 0
from where
(z ; 1)2 xz 2 + (y + 2x)z + x = 0:
We have one of the following possibilities:
(i) The number z = 1 is the only to satisfy the equality xz 2 + (y + 2x)z + x = 0. This is true
if x + y + 2x + x = 0 and D = (y + 2x)2 ; 4x2 = y (y + 4x) = 0, from where y = ;4x and y > 0.
(ii) There is no real z , which satisfy the equality xz 2 + (y + 2x)z + x = 0. Then, D =
(y + 2x)2 ; 4x2 = y (y + 4x) < 0, from where y < ;4x and y > 0.
Therefore, every time when y ;4x and y > 0, the point (x y ) belongs to the locus.
Case 2. Let z = ;1. Analogously to the previous case we nd that y < 0 and then the
given equality takes the form
(z + 1)2 xz 2 + (y ; 2x)z + x = 0:
We have the following possibilities:
(i) the number z = ;1 is double root of the equation
xz2 + (y ; 2x)z + x = 0: ()
(ii) the equation (*) has no real root.
By computing we deduce that here y 4x and y < 0.
5
Finally, the locus (Figure 11) consists of the internal as well as of the boundary points of the
angle which is dened by the graphs of the linear functions y = 4x and y = ;4x when x < 0,
without the points from the negative part of the x { axis.
Problem 2. In the triangle ABC , ha and hb are the altitudes from A and B respectively,
`c is the internal bisector of 6 ACB, while O, I and H are the circumcenter, the incenter and
the altitude center, respectively. Prove that if `c + `c = 2, then OI = IH .
ha hb
Solution. Let BC = a, AC = b, 6 ACB = , CL = `c Figure 12.
1
(Figure 12). Since SABC = SALC + SBLC , then 2 `c a sin 2 +
1 ` b sin = 1 ab sin , from where ` = 2ab cos 2 . But
2 c 2 2 c a+b
2a cos
ha = b sin and hb = a sin . Then h`c + h`c = (a + b) sin2 +
a b
2b cos 2 1 a + b 1 , which
= =
(a + b) sin sin 2 a + b a + b sin 2
shows that sin 2 = 12 , i.e. 2 = 30 , because 2 < 90. Conse-
quently, = 60.
Let CL meet the circumcircle of 4ABC in M . Then
OM ? AB. But CH ? AB, and therefore 6 OMC = 6 MCH . On the other hand OM = OC =
R (R is the circumradius). Consequently, 6 OMC = 6 OCM = 6 MCH . Let H2 be the foot of
the altitude from B . From 4HCH2 we have CH = CH sin
2
(because 6 H2 HC = 90 ; 6 ACH
= 6 CAB = ). But CH2 = a cos , i.e. CH = a cos = 2R cos . Since = 60 , then
sin
CH = R.
We consider 4COI and 4CHI . Since the point I lies on CM , then 6 OCI = 6 HCI . Also,
CO = R = CH and consequently 4COI
= 4CHI . Thus, OI = IH .
Problem 3. Let A1, A2, : : :, An (n 4) be n points in the plane, no 3 of which are colinear.
a) Prove that there is at most one point As , such that all triangles As Ai Aj (i j = 1 2 : : : n)
are acute.
b) Let among A1, A2 , : : :, An be a point, which is a vertex of an acute triangles only. We
consider the angles dened by the given points. Denote by Nk the number of the acute angles
6 Ai Ak Aj (i j = 1 2 : : : n) for which the point Ak is their vertex. Find the minimal value of
Nk .
Solution. a) Assume that there is more than one point with the given property and let A
and B be such two points, while X and Y be any two of the remaining points. There are two
possibilities for the points A, B , X and Y :
(i) A, B , X and Y dene a convex quadrilateral. Because the sum of its internal angles is
360 , then at least one of these angles is not acute. Consequently, at least one of the triangles
6
of acute triangles only.
b) Let A1 be the point which is a vertex of acute triangles only. We consider the angles
6 Ak A1 As (Figure 13). Let A2 A1 An be the biggest one. Because all angles with vertex A1 are
acute, the points A3 , A4 , : : :, An;1 are inside the acute angle 6 A2 A1An . We can assume that
the points are enumerate in such a way that 6 A2A1 Ak < 6 A2A1 Ak+1 for k = 3 4 : : : n.
Consider 6 Ai Ak Aj , where 2 i < k < j n. Figure 13.
We assume that Ak is internal for 4A1 Ai Aj . Then
6A1Ak Ai + 6 A1Ak Aj > 180
and at least one of the angles 6 A1 Ak Ai and 6 A1Ak Aj will not be
acute, which is a contradiction.
Therefore, the points A1 , Ai , Ak , Aj dene a convex quadrilat-
eral. If we assume that 6 Ai Ak Aj 90, then there exists an angle
of this quadrilateral with vertex A1 , Ai or Aj which is 90. This
is impossible.
Therefore, 6 Ai Ak Aj > 90 , and it is clear that 6 Ai Ak Aj < 90 when i j < k or i j > k.
Particularly, it follows from here that no of the angles Ai Ak Aj is right. The number of all
angles with vertex Ak is equal to (n ; 1)(2 n ; 2) . If Tk is the number of the obtuse angles with
vertex Ak , then Nk = (n ; 1)(2 n ; 2) ; Tk . Thus, Nk is minimal when Tk is maximal.
It is easy to see that T1 = T2 = Tn = 0. Let 3 k n ; 1. It is clear that the number of
the points Ai for which 2 i < k is k ; 2, and the number of the points Aj for which k < j n
is n ; k. Then Tk = (k ; 2)(n ; k) for k = 3 4 : : : n ; 1.
But (k ; 2)(n ; k) (n ;4 2) , and the equality is reached when k ; 2 = n ; k, i.e. when
2
2 4
(ii) The number n is odd. Then, the nearest integers to n +2 2 are n +2 1 and n +2 3 . It is
easy to see that now the maximal value of Tk is T n+1 = T n+3 =
n + 1 ; 2 n ; n + 1 =
2 2 2 2
(n ; 1)(n ; 3) . Consequently, the minimal value of N in this case is
4 k
= (n ; 1)(n ; 2) ; (n ; 1)(n ; 3) = (n ; 1) :
2
N n+1
2
= N n+3
2 2 4 4
Grade 11
Problem 1. Find the values of the real parameter a, for which the inequality x6 ; 6x5 +
12x4 + ax3 + 12x2 ; 6x + 1 0 is satised for all real x.
7
Solution. When x = 0 the given inequality
is satised
for all a. Thus, it is enough to nd
3 1
such a, that x + 3 ; 6 x + 2 + 12 x + + a 0 for all x > 0 and x3 + 13 ;
2 1 1
x x x x
6 x +
2 1 1
+ 12 x + +a 0 for all x < 0.
x2 x
Denote t = x + x1 . It is clear that x > 0 () t 2 and x < 0 () t ;2. But
x2 + x12 = t2 ; 2 and x3 + x13 = t3 ; 3t.
We consider the function f (t) = t3 ; 3t ; 6 t2 ; 2 + 12t+ = t3 ; 6t2 + 9t + 12 + a.
The problem is reduced to nd such a, that f (t) 0 for all t 2 and f (t) 0 for all t ;2
simultaneously.
Figure 14. But f 0 (t) = 3t2 ; 12t + 9 = 3(t ; 1)(t ; 3), from where it is
easy to obtain that (Figure 14) f (t) 0 for all t 2 i f (3) 0
and f (t) 0 for all t ;2 i f (;2) 0. Since f (3) = a + 12
and f (;2) = a ; 38, we nd for a: a ;12 and a 38. Finally,
;12 a 38. _
Problem 2. The point D lies on the arc BC of the circumcircle
of 4ABC which does not contain the point A and D 6= B , D 6= C .
On the rays BD! and CD! there are taken points E and F , such that BE = AC and CF = AB .
Let M be the midpoint of the segment EF .
a) Prove that 6 BMC is right. _
b) Find the locus of the points M , when D describes the arc BC .
Figure 15. Figure 16.
8
1 ;!2 ;!2 ;! ;! 2
= 4 CF + BE + 2 CF BE ; a
= 14 c2 + b2 + 2bc cos( ; ) ; a2
= 14 c2 + b2 ; 2bc cos ; a2 = 0:
(We have used that jBE j = jAC j = b, jCF j = jAB j = c and 6 (;! BE ;!
CF ) = 6 BDC = ; .)
;;! ;;!
Consequently, BM ? CM , i.e. the angle 6 BMC is right.
b) It is clear that the points M and A are in dierent semiplanes with respect to the line BC .
According to a) 6 BMC is a right angle and consequently if the point M is from the locus, then
M lies on a semicircle k with diameter BC , k and A are in dierent semiplanes
with respect to
the line BC . Let BCM = . Then CM = a cos (because BMC = 2 ) and
6 6
;;! ;!
CM CB = a(a cos ) cos = (a cos )2:
On the other hand by the sine theorem we nd:
CM CB = 21 ;!
;;! ;!
CB + ;!BE + ;!CF ;! CB
= 12 a2 + ab cos ( ; ( ; ')) + ac cos '
1 b c
= 2 a 1 ; a cos( ; ') + a cos '
2
9
Spring mathematics
tournament|1997
Problem 8'1. Given the equation jx ; aj + 15 = 6jx + 2j, where a
is a real parameter.
(a) Prove that for any value of a the equation has exactly two
distinct roots x1 and x2.
(b) Prove that jx1 ; x2j 6 and nd all values of a for which
jx1 ; x2j = 6.
Solution: (a) When x ;2, the equation is equivalent to jx ;
aj = ;6x ; 27, which has a solution only if ;6x ; 27 0, i. e., if
x ; 29 . Considering the cases of both x a and x a shows that
if x ;2, the given equation has a unique root x1, and if x > ;2,
it has a unique root x2:
8 a ; 27 9 8 3 ; a a 1
>< a < ;2 ><
7
x1 = > a + 27 9 x2 = > a +5 3 a > 12
:; a ; :
5 2 7 2
1
(b) It follows from (a) that
8 156 ; 12a 9
>>
>< 35 a < ; 2
jx1 ; x2j = > 6 ; 29 a 12
>> 12a + 204 1
: a >
35 2
It remains to be seen that 156 ; 12a > 6 when a < ; 9 and
35 2
12a + 204 > 6 when a > 1 . Therefore jx ; x j 6 and equal-
35 2 1 2
2
Problem 8'.3. Find all natural numbers n such that there exists
an integer number x for which 499(1997n + 1) = x2+x.
Problem 8.1. Find all values of the real parameter m such that
the equation (x2 ; 2mx ; 4(m2 + 1))(x2 ; 4x ; 2m(m2 + 1)) = 0 has
exactly three distinct roots.
3
roots. Thus D2 = 4 + 2m(m2 + 1) = 0 and so m = ;1. But this
case has already been considered. Thus we determine that m = 3.
5
Solution: Let M and m be the maximum and minimum values
of f (x) in the interval 0 1]. Then the condition of the problem is
equivalent to M 1 and m ;1. There are three cases to consider.
Case 1: a 2 0 1]. Then m = f (a) = ;2a2 ; 43 and M = f (0) =
;a2 ; 43 or M = f (1) = ;a2 ; 2a + 14 . It follows from m ;1 and
p p
M 1 that a 2 ; 42 42 ] and a 2 (;1 ; 32 ] ; 12 1). In this
p
case the solution is a 2 0 42 ].
Case 2: a < 0. Now m = f (0) and M = f (1). From m ;1
and M 1 we get that a 2 ; 12 12 ]: Therefore a 2 ; 21 0).
Case 3: a > 1. Now m = f (1) and M = f (0). It follows from
m ;1 and M 1 that a 2 ; 25 12 ] which is a contradiction with
a > 1. p
1
Thus the solution is a 2 ; 2 42 ].
6
Solution: (a) We shall use the usual notation for a triangle. Let
CL and CM be respectively the bisector and the median from C .
It follows from CG = 32 CM that SCIG = 32 SCIM . Thus SCIM =
SCLM ; SILM = LM2 hc ; LM2 r = LM 2 (hc ; r). We nd from
AL + BL = c and BL AL = b that AL = bc and so LM = AM ;
a a+b
c bc c ( a ;
AL = 2 ; a + b = 2(a + b) . Also, hc ; r = 2cS ; r = 2pr
b )
c ;r =
r(2p ; c) = r(a + b) . Therefore S = 2 S = 2 c(a ; b)
CIG
c c 3 CIM 3 4(a + b)
r(a + b) = (a ; b)r .
c 6
(b) The distances from I and G to AB are respectively r and
hc . Hence r = S = chc = chc = hc and so IGkAB . Therefore
3 p 2p 3c 3
the altitude from C of triangle CIG equals 23 hc = 2r. Thus SCIG =
IG r. On the other hand SCIG = (a ;6 b)r = r3 and so IG = 13 .
Problem 10.1. Find the least natural 3number a such that the
equation cos (a ; x) ; 2 cos (a ; x)+cos 2a cos( x
2 x +
2a 3
)+2 = 0
has a root.
Solution: The roots of the our equations are the common roots of
cos (x ; a) = 1 and cos 32x
a cos( x + ) + 1 = 0. The roots of
2a 3
the rst one are x = a + 2n n = 0
1
2 : : : and the roots of the
second one are x = 2a(k ; 1 ) k = 0
1 : : : : Therefore a = 6n
3 6k ; 5
8
for some integer numbers n and k. It is easy to see now that the
least natural number with the required property is a = 6.
10
number of diagonals from each vertex. Therefore both Aj;1As and
Aj+1As are diagonals so Aj;1Aj+1As is of type t3. Hence there is a
triangle of type t3 adjacent to each triangle of type t1. If distinct
triangles of type t1 are adjacent to distinct triangles of type t3 then
x1 x3 = x1 ; 2 < x1, a contradiction. Therefore there are at
least two triangles of type t1 adjacent to one and the same triangle
of type t3. Without loss of generality assume these are the triangles
A1AnAn;1 and An;1An;2An;3 . Consider the polygon A1A2 An;3.
Obviously the diagonals partition this polygon into triangles and
there is an even number of diagonals through each vertex.
Conversely, if the polygon A1A2 An;3 can be partitioned in the
required way, then adding the vertices An;2, An;1 , An and diagonals
An;3An;1 and A1An;1 shows that the same is true for the polygon
A1A2 An.
Therefore a natural number n 6 is a solution if and only if
n ; 3 is a solution. It is easy to see that n = 3 is a solution, whereas
n = 4 and n = 5 are not. Thus all natural numbers satisfying the
conditions of the problem are n = 3k, k = 1 2 : : :.
Problem 11.1. For any real number b denote by f (b) the maximal
value of j sin x + 3 + 2sin x + bj. Find the minimal value of f (b).
13
Spring mathematics
tournament { 1998
Problem 8'.1. Find all values of the real parameter a such that
the inequalities jx + 1j + j2 ; xj < a and 6x 5;a 5;a 8+ 5 < ; 12 are
equivalent.
1
Let us rewrite the second inequality in the form 66xx+;55aa;+115 <
0. Its solutions form either the interval 11 ;6 5a
5a 6; 5 or the
5a ; 5 and 11 ; 5a is greater.
6 6
Therefore the two inequalities are equivalent if
1 ; a = 11 ; 5a 1 + a = 5a ; 5
2 6 2 6
or
1 ; a = 5a ; 5 1 + a = 11 ; 5a :
2 6 2 6
The rst pair is satis ed by a = 4, the second one by a = 1. But if
a = 1, the rst inequality has no solution, whereas the second one
does. Thus the only solution is a = 4.
2
of the problem implies that
MP is the bisector of NMK 6
= 4PKM
and PNM = PKM =
90 deg ; 2 . On the other hand,
6 6
AN = BM . Therefore AC = BC .
3
2. Let H lie between N
and P and let K be a point
such that MK = MN and
HP = HK . It follows from
the isoscelesness of 4KHP
that HPK = HKP = 2 .
6 6
AC = BC = 2. Then SABC = 21 AC BH = 1.
2. When = 90 deg : AC ? BC , AC = BC = 2BH = 2.
Then SABC = 12 AC BC = 2.
4
are exactly 1998 intersecting points?
Problem 8.2. Let l1 and l2 be the loci of the centroid G and the
incentre I of the right triangle ABC whose hypotenuse AB is a given
segment of length c.
(a) Find l1 and l2.
6
(b) Find the area of 4ABC when the length of GI is minimal.
circle k2 with centre Q and radius R and the other one is part of a
0 0
circle k2 with centre Q and the same radius and Q and Q lie in
00 00 0 00
which does not lie in the same semiplane as C does with respect to
AB . Without loss of generality this is Q . Denote the intersecting
0
BI is the bisector of ABC and C lies in one and the same semiplane
6
c deg, it follows
6
7
N , and l2 consists of two arcs k2 and k2 such that if X 2 k2 or
0 00 0
(b) We shall show that R > 23 c and it will follow (how?) that
k1 lies in the interior of both k2 and k2 . Let P be the intersecting
0 00
Since PQ = 3c = 6c + 6c = 0
R = OQ + OG + GI and so
0
G1 I1 > GI .
8
Case 2: G2 2 l1 , I2 2 l2
and O are not on a line. Let
OI2 intersect k1 in a point G1.
Then G2I2 > G1 I2. Indeed,
it follows from 4OG2I2 that
OG2 + G2I2 > OI2. But OI2 =
OG1 + G1I2 = OG2 + G1I2 and
therefore G2I2 > G1I2. But
from case 1 we get that G1I2 >
GI and so G2I2 > GI .
Thus 4ABC having the required property is such that Q , O, G
0
and so SABC = c4 .
9
Use induction by n. It is easy to see that the above formula holds
for n = 2. Suppose it is true for some n. To obtain the result we
count how many new regions are added when a new point appears on
the circle. It is easily seen that if a chord intersects t other chords,
then it `adds' t + 1 new regions. Therefore the new regions are:
nX1
(k(n ; k ; 1) + 1):
;
k=0
121 chords such that there are 1998 regions. Let us draw all chords
with ends in 16 of the given points (there are 120 such chords). !
It follows then that the interior of the circle is divided into 164 +
!
16 + 1 = 1941 regions. Draw a chord connecting the 17th point
2
with one of the rst 16 in a way that there are 8 and 7 points on the
two sides of the drawn chord. This chord intersects 8 7 = 56 chords
and therefore there are 57 new regions. Therefore the total number
of regions is 1941 + 57 = 1998.
10
Problem 9.1. Find all parameters a such that the inequality jax2;
3x ; 4j 5 ; 3x holds for any x 2 ;1
1].
Solution:
Let E be the second in-
tersecting point of PB and
the circle. Then PA2 =
PE PB ) PD = PE .
Hence APD = EPC . (If
6 6
BP CP BC
AP : (1)
BP
Also, 4APB 4DCP because APB = DPC = APC ;
6 6 6
AP = BP and so AB = BP : (2)
APD and DP CP DC CP
6
12
From (1) and (2) we get BC AB AD = AP BP = 1. On the
DC BP CP
other hand,
SABD = 21 AB AD sin DAB = AB AD = 1
6
Problem 10.1. Find all values of the real parameter a such that
the inequality x4 + ax3 + (a + 3)x2 + ax + 1 > 0 holds for all real
values of x.
15
same property holds for 2 4 8 and 16, we will be done. Assume 2k
divides a2b + 1, where 1 k 3. Then a is an odd integer number
and therefore a2 ; 1 is divisible by 8. Consequently b + 1 is divisible
by 2k and thus a2 + b is divisible by 2k as well. Assume 16 divides
a2b +1. Then a is an odd integer number and it is easy to verify that
a2 is congruent to 1 or 9 modulo 16. Further, b should be congruent
to 15 or 7 modulo 16 correspondingly and again a2 + b is divisible
by 16.
The required numbers are all divisors of 240.
Problem 11.1.
a.) Let p be a positive real parameter. Find the least values of
the functions f (x) = x + xp and g(x) = x + xp2 in the interval
(0
+1).
Let a1 ap2 a3 be positive real numbers. Prove that 3(a1 +
b.) p
a1a2 + 3 a1a2a3) 4(a1 + a2 + a3):
p (x ; pp)(x + pp)
Solution: a.) Since f (x) = 1 ; x2 =
0
xp2 , we get
that the functionpf (x) decreases in the interval (0
p) and increases
in the interval ( p p
+1). pTherefore the minimal value of f (x) in
(0
+1) equals f ( p) = 2 p. Analogously it follows from g (px) = 0
1; 2xp3 that the least value of g(x) in (0
+1) equals g( 3 2p) = 3 22p .
q 3
16
4 1 4 y 2 1 4 z 3 1
that 3 y 3 x + 3 x and 3 y + 3 y2
z. Therefore 3 x + 3 x ; 4 y 2
y + 43 zy2 ; z
13 y + 43 zy2 ; z
0, which completes the proof. Note
3 3
Problem 11.2. Let I and r are the incentre and inradius of 4ABC ,
and N is the midpoint of the median through C . Prove that if
r = CN ; IN , then AC = BC or ACB = 90 deg :
6
18
Spring mathematics
tournament|Kazanl^ak,
30 March{1 April 1999
Problem 8'1. Given an inequality jx ; 1j < ax, where a is a real
parameter:
a) Solve the inequality.
b) Find all values of a such that the inequality has exactly two
integer solutions.
Chavdar Lozanov, Kiril Bankov, Teodosi Vitanov
1 x.
b) It follows from a) that the inequality could have two integer
solutions only if 0 < a < 1. Since in this case 0 < a +1 1 < 1 < 1 ;1 a ,
we nd that there are exactly two integer solutions if and only if
2 < 1 ;1 a 3:
3
Prove that there exists a rectangle with sides parallel to the coordi-
nate axes whose vertices are all of the same colour.
Kiril Bankov
9
that 6 Q3Q1Q2 = 6 A3A1A2, and the same equalities for the remain-
ing pairs of angles. Thus 4P1P2P3
4Q1Q2Q3
4A1A2A3 and
since the rst two triangles have the same circumcircle, they are
identical.
b) Since A1P1KQ1 is a parallelogram, we obtain 6 A1P1Q1 =
^ ^
6 KQ1 P1 = P2 Q2 + P1 Q2 = 6 P1 P2 P3 = 6 A1A2 A3. Similarly 6 A1Q1P1 =
6 A1 A2A3.
2 2
Let O be the circumcentre of 4A1A2A3. It is easy to see that
OA1 ? P1Q1,OA2 ? P2Q2 and OA3 ? P3Q3. Denote the midpoint
of OK by S . If R1 R2 and R3 are the midpoints of P1Q1 P2Q2 and
P3Q3 then SR1 k OA1
SR2 k OA2
SR3 k OA3 (SR1 is a middle
segment in 4OKA1 =) SR1 k OA1). Therefore S lies on the axes
of symmetry of P1Q1
P2Q2 and P3 Q3, so S is the centre of the circle
through the points P1 P2 P3 Q1 Q2 Q3.
10
It is easy to nd now that p + q = 3 and so p = 1 q = 2 or
p = 2 q = 1. In the rst case we get that f (x) = x3 + x2 ; x ; 1,
which is a solution and in the second one f (x) = x3 ; x2 ; x + 1,
which is not.
Second case: ja0j 2. Now 0 = f (a0) = jan0 + an;1an0 ;1 n+;1 +
a1a0+a0j ja0jn ;ja0jn;1; ja0j2;ja0j;ja0j = ja0j(ja0j;ja2)(0j;ja10j ;1)
0. Therefore ja0j = 2. Moreover an;1an0 ;1 : : : a1a0 a0 have the same
negativity|the opposite to those of an0 .
We conclude that a0 = ;2 n is an even number and ai = (;1)i+1
for i = 1 2 : : : n ; 1. If n > 2, then a2 = ;1 and 0 = f (;1) =
(;1)n + (;1)n (;1)n;1 + + (;1)2(;1) ; 2 = ;n ; 2 6= 0, which
is impossible. Therefore n = 2 and f (x) = x2 + x ; 2, which is a
solution.
Answer:
f (x) = x2 + x ; 2 and f (x) = x3 + x2 ; x ; 1:
+x+1
get a contradiction.
4C1D1M
4DAM
4D1A1M , we obtain
AB = BM
BC = BM
CD = DM
DA = DM :
A1B1 A1M B1C1 C1M C1D1 C1M D1 A1 A1M
12
It is easy to see now that
AB CD = BM DM = BC DA :
A1B1 C1D1 A1M C1M B1C1 D1 A1
Using that AB = BC = CD = DA it follows from the above that
A1B1 C1D1 = A1D1 B1C1, Q. E. D.
2 2
This is a contradiction with the number of drawn segments.
Therefore the answer is 3n 2; 3 .
Problem 11.1. Given a function f (x) dened for any real x and
f (tgx) = sin 2x for x 2 (; 2 2 ). Find the minimum and the maxi-
mum of the function f (sin3 x) f (cos3 x).
Oleg Mushkarov, Nikolai Nikolov
14
Let u = sin x cos x = 21 sin 2x. Then u 2 ; 21 12 ] and we have to nd
the minimum and the maximum of the function g(u) = 2 ; 34uu3 + u6
3
15
the above becomes
cos2 A2 = 2R sin A2 (sin A2 + cos B ;2 C ) = 2R sin A2 cos B2 cos C2 :
s s s
16
number n denote the square with vertices (n n) (;n n) (;n ;n)
and (n ;n) by Kn .
Since A is a nite set, there exists a natural number d such that
A Kd . For every n denote the number of arrows within the square
Kn by Sn . Since there are at least s + 1 arrows pointing out of every
integer point of Kn (and these arrows are within the square Kn+d ),
it follows that (2n + 1)2(s + 1) Sn+d . On the other hand, since AX
is a central-symmetric set, there are at most s ; 1 arrows pointing to
every integer point of Kn+d . Therefore Sn+d (2n + 2d + 1)2(s ; 1).
Thus (2n + 1)2(s + 1) (2n + 2d + 1)2(s ; 1), so s + 1 (1 +
2d )2(s ; 1) for any n. When n ! 1 one obtains s + 1 s ; 1,
2n + 1
a contradiction. Therefore a point X with the required property
does exist.
17
WINTER MATHEMATICAL COMPETITION
1995
Solution. Let F be the midpoint of the segment AD (Figure 1). Then the quadrilateral
MPNF is a parallelogram. Hence K is midpoint of the segment FP . It follows from here that
KL = 21 :FQ. On the other hand FQ = 21 :a (because F and Q are midpoints of ED and AD
respectively).
Therefore KL = 14 :a.
1
Problem 3. Every point in the plane is colored either in black or in white. Prove that
there exists a right angled triangle with hypotenuse of length 2 and an acute angle of 60 , which
vertices are colored in one and the same colour.
Solution. First we shall show that there exist two points which are colored in one and the
same colour and the distance between them is 2. Indeed let ABC be a equilateral triangle of
side 2. Obviously two of its vertices (say A and B ) are colored in one and the same colour (e.g.
white). Let AXY BZT be a regular hexagon with a big diagonal AB (Figure 2).
If one of the vertices X , Y , Z , T (e.g. X ) is white, then the vertices of the triangle ABX
(6 AXB = 90 , 6 XAB = 60 ) are colored in one and the same colour. Otherwise the vertices
of the triangle XY T (6 Y XT = 90 , 6 XY T = 60 ) are colored in one and the same colour.
Grade 8.
2
that the quadrilateral DEMN is inscribed in a circle if and only if there exists a circle, inscribed
in the quadrilateral CMGN .
Figure 3. Figure 4.
Solution. Without loss of generality we can suppose that N is between A and E . There are
two possibilities for the points D and E which are shown in the Figure 3 and Figure 4.
The quadrilateral DEMN is inscribed in a circle if and only if 6 CNM = 6 CDE , i.e.
6 BAN + 6 BDE = 1800 (because MN kAB and 6 CNM = 6 BAN ). Thus the quadrilateral
DEMN is inscribed in a circle i the quadrilateral ABDE is inscribed in a circle. This is
equivalent to 6 DAE = 6 DBE , i.e. to AC = BC .
Therefore we should prove that there exists a circle, inscribed in the quadrilateral CMGN
if and only if AC = BC .
Let AC = BC . Then CM = CN and since G is the cen- Figure 5.
ter of gravity of the triangle ABC , we have GM = 3 AM = 1
1 BN = GN . Hence CM + GN = CN + GM , i.e. there ex-
3
ists a circle, inscribed in the quadrilateral CMGN (Figure
5).
Conversely if there exists a circle, inscribed in the quadri-
lateral CMGN , then CM + GN = CN + GM and 12 BC +
1 BN = 1 AC + 1 AM . Hence AM ; BN = 3 (BC ; AC ).
3 2 3 2
Let K and L be the points of contact of the circle and the
sides CM and CN respectively. Obviously CK = CL.
On the other hand this circle is inscribed in the triangles
ACM and BCN . Hence CK = 21 (AC + CM ; AM ), CL =
1 (BC + CN ; BN ). Thus AM ; BN = 1 (AC ; BC ). It follows from here that 3 (BC ; AC ) =
2 2 2
1 (AC ; BC ), i.e. AC = BC .
2
Problem 3. Thirty points are given in the plane. Some of them are connected with segments
as it is shown in the Figure 6. The points are labeled with dierent positive integers.
If a is a segment and p and q are the numbers, corresponding to its endpoints, we denote
(a) = jp ; qj.
a) Construct an example of labeling of the points with the integers 1 2 : : : 30, in which there
3
exists exactly one segment a with (a) = 5
b) Prove that for every labeling there exists at least one segment a with (a) 5.
Figure 6. Figure 7.
1 ; x1 1 ; x2 9
Solution. a) Since x1 and x2 are real numbers, then
D (f ) = (m ; 4)2 ; 4 m2 ; 3m + 3 = ;3m2 + 4m + 4 0:
Hence ; 23 m 2. On the other hand
6 = x21 + x22 = (x1 + x2 )2 ; 2x1x2 = ;m2 ; 2m + 10:
p
We obtain from here that m = ;1 5. But
p p
;1 ; 5 < ; 32 < ;1 + 5 < 2
4
p
and therefore only m = 5 ; 1 is a solution of the given problem.
b) We have
mx21 + mx22 = mx21(1 ; x2) + x22(1 ; x1 )]
1 ; x1 1 ; x2 f (1)
= x1 + x2 ; x1x2 (x1 + x2 )
2 2
m;2
= m ; 8m + 13m ; 2 = m2 ; 6m + 1
3 2
m;2
Thus if F = 1mx mx2 + 8, then F = (m ; 3)2 and
2 2
1
+
; x1 1 ; x2
121 = ; 2 ; 32 F > (2 ; 3)2 = 1:
9 3
Problem 2. The point D lies inside the acute triangle ABC . Three of the circumscribed
circles of the triangles ABC , ABD, BCD and CAD have equal radii. Prove that the fourth
circle has the same radius.
Solution. There are two cases:
1. The radii of the circumscribed circles of the Figure 8.
triangles ABD, BCD, CAD are equal.
Let O1 , O2 and O3 be the centers of these cir-
cles (Figure 8). Obviously the quadrilaterals O2 CO3D,
O3 AO1D and O1BO2 D are rhombuses.
Hence O2C kO3DkAO1 and O2 C = AO1 . Thus the
quadrilateral CAO1O2 is a parallelogram. It follows
from here that AC kO1O2 and since O1 O2 ? DB , then
BD ? AC . Analogously CD ? AB and AD ? BC .
Therefore D is the altitude center of 4ABC .
Now it is easy to see that 6 BDC = 1800 ; 6 BAC ,
which implies that the circumscribed circles of 4BCD
and 4ABC are symmetric according to the line BC .
Therefore their radii are equal.
2. The radii of the circumscribed circles of the triangles ABC , ACD and BCD are equal.
In this case the circumscribed circles of 4ABC and 4BCD are symmetric according the
line BC . Hence 6 BDC = 1800 ; 6 BAC . Analogously 6 ADC = 1800 ; 6 ABC . Therefore
6 ADB = 3600 ; 6 BDC ; 6 ADC = 1800 ; 6 ACB and the circumscribed circles of 4ABD and
4ABC are symmetric according to the line AB. Thus they have equal radii.
Problem 3. Let A be a set with 8 elements. Find the maximal number of 3-element subsets
of A, such that the intersection of any two of them is not a 2-element set.
Solution. Let B1 B2 : : : Bn be subsets of A such that jBi j = 3, jBi \ Bj j 6= 2 (i j = 1 : : : n).
Assume that there exists an element a 2 A, which belongs to four of the subsets B1 B2 : : : Bn
(e.g. a 2 B1 B2 B3 B4). Then jBi \ Bj j 1 (i j = 1 : : : 4). But Bi = 6 Bj if i 6= j , i.e.
jBi \ Bj j 6= 3. Thus jBi \ Bj j = 1 (i j = 1 2 3 4). It follows from here that jAj 1 + 4:2 = 9,
5
which is a contradiction. Therefore every element of A belongs to at most three of the subsets
B1 B2 : : : Bn . Then 3n 8:3, i.e. n 8.
If A = fa1 a2 : : : a8g, then the subsets
B1 = fa1 a2 a3g B2 = fa1 a4 a5g B3 = fa1 a6 a7g B4 = fa8 a3 a4g
B5 = fa8 a2 a6g B6 = fa8 a5 a7g B7 = fa3 a5 a6g B8 = fa2 a4 a7g
provide an example of exactly eight 3-element subsets of A, such that jBi \ Bj j 6= 2.
Therefore the searched number is n = 8.
Grade 10.
Problem 1. Find all positive roots of the equation
log(x+a;1) x +4 1 = loga 2
where a > 1 is a real number.
Solution. It is clear that if a > 1 and x > 0, then x + a ; 1 > 0 and 4 > 0. Hence in
x+1
this case log(x+a;1) 4
x + 1 is well dened. Since
log 4
log(x+a;1) 4 =
a x+1
x + 1 loga (x + a ; 1)
then the given equation is equivalent to
4 = 2loga(x+a;1) : (1)
x+1
The function x +4 1 is strictly decreasing in the interval (0 +1). The function loga (x + a ; 1)
is strictly increasing in the interval (0 +1) and obviously the same is true for the function
2loga (x+a;1) . Therefore the equation (1) has no more than one root in the interval (0 +1).
On the other hand it is easy to check that x = 1 is a root of this equation.
Problem 2. A circle _
k with center O and diameter AB is given. The points C and D are
moving along the arc AB so that C is between B and D and if 6 BOC = 2 and 6 AOD = 2,
then tan = tan + 32 . Prove that the lines, which are perpendicular to CD and divide CD in
ratio 1 : 4 measured from C , pass through a xed _
point of the given circle.
Solution. Let E be such a point on the arc AB , not containing C and D, that if 6 BOE = 2 ,
then tan = 2 (Figure 9).
We shall show that the point E satises the problem's conditions.
It is enough to prove that if F is the foot of the perpendicular from E to CD, then the point
F is between C and D and FD CF = 1 .
4
We have
tan(6 ECD) = tan(6 ECA + 6 ACD) = tan(
2 ; + )
6
tan
2 ; + tan 1 + tan
tan
=
= 2 tan = 12+;2tan
1 ; tan 2 ; : tan 1 ; 2
3
1 + 2 tan + 2
= 3 = 41+ 2 tan = 8 + 4 tan
1 ; 2 tan
2 ; tan + 2 2 ; tan
tan(6 EDC ) = tan(6 EDB + 6 BDC ) = tan( + ) = 1tan + tan = 2 + tan :
; tan : tan 1 ; 2 tan
Thus Figure 9.
tan(6 ECD) = 4 tan(6 EDC ):
It follows from the last equality that either
6 ECD
2 and 6 EDC
2
or
6 ECD <
2 and 6 EDC <
2 :
But these angles belong to the triangle EDC .
Hence they are acute angles. Therefore the point
F is between the points C and D. Since
tan(6 ECD) = EF EF
FC and tan( EDC ) = FD
6
then EF =
FC FD 4 EF , i.e. FD = 4:FC .
Problem 3. Find all prime numbers p, for which the number p(2p;1 ; 1) is a k-th power
(k > 1) of a positive integer.
Solution. Let p(2p;1 ; 1) = xk (x > 0 is an integer). It is clear that p 6= 2, i.e. p = 2q + 1 is
an odd number. Since p=x, then x = p:y (y is a positive integer) and (2q ; 1)(2q + 1) = pk;1 y k .
At least one of the numbers 2q ; 1 and 2q + 1 is a k-th power of an integer, because they are
relatively prime numbers.
1. Let 2q ; 1 = z k , i.e. 2q = z k + 1. If k is even, then z k + 1 is not divisible by 4 . Hence
q = 1, p = 3 and p(2p;1 ; 1) = 32.
If k = 2l + 1 then 2q = (z + 1)(z 2l ; z 2l;1 + ; z + 1), i.e. z + 1 = 2, where 0 < q .
On the other hand
2q = (2 ; 1)2l+1 + 1 = 22:A + 2 (2l + 1)
(A is an integer). The last equality contradicts with < q .
2. Let 2q + 1 = z k , i.e. 2q = z k ; 1. If k is odd, we obtain a contradiction as in the previous
case.
If k = 2l, then (z l ; 1)(z l + 1) = 2q and since GCD(z l ; 1 z l + 1) = 2, we have z l ; 1 = 2,
i.e. q = 3, p = 7, p(2p;1 ; 1) = 7:63 = 212.
The sought numbers are p = 3 and p = 7.
7
Grade 11.
Problem 1. Find all the values of the real parameter p, for which the range of the function
f (x) = 2(1p;;psin
) + cos x :
2x
g(y) = ky2 ; y ; q(k + 2) has as its rst coordinate y 0 = 21k 2 (0 1] ;1 1]. Therefore the
equation g (y ) = 0 has at least one root y0 2 ;1 1] i at least one of the inequalities g (;1) 0
and g (1) 0 holds. It is easy to obtain from here that q kk + + 1 for every k 2 1 2]. Hence
2
2 1 2 1
q 3 , i.e. q 2 ; 32 3 and q 6= 4 .
1 33
Finally p 2 3 32 , p 6= 43 .
Problem 2. The point O is on the edge AB of the tetrahedron ABCD. The circumscribed
sphere of the tetrahedron AOCD intersects the edges BC and BD in the points M and N
(M 6= C , N 6= D) respectively. The circumscribed sphere of the tetrahedron BOCD intersects
the edges AC and AD in the points P and Q (P 6= C , Q 6= D) respectively. Prove that the
triangles OMN and OQP are similar.
Solution. Since the points A, C , M and O are in one and the same plane and are lying on
a sphere, then the quadrilateral ACMO is inscribed in a circle (Figure 10). Then 4BOM
4BCA, and OM BM OQ AQ
CA = BA (Figure 11). Analogously from 4ABD: BD = AB . Hence
OM = BM AC : (1)
OQ AQ BD
8
From 4BCD we have MN =
CD BD
BM , and from 4ACD: PQ = AQ . Thus
CD AC
Figure 10. Figure 11.
MN = BM AC : (2)
PQ AQ BD
It follows from (1) and (2) that OM MN ON MN
OQ = PQ . Similarly OP = PQ .
Therefore 4OMN
4OQP .
Problem 3. Solve in positive integers the equation:
1 + 5x = 2y + 2z :5t:
Solution. If y 2 and z 2, then the right side of the given equation is divisible by 4. But
1 + 5x 2 (mod 4) and hence min(y z ) = 1. On the other hand 2y 1 (mod 5). Thus y is
divisible by 4 (4 is the index of 2 modulo 5). It follows from here that y 4, z = min(y z ) = 1
and the equation is 1 + 5x = 24y0 + 2:5t, where y = 4y0 (y0 is a positive integer).
If t = 1 then 5x ; 24y0 = 9. Using congruence modulo 3 we obtain that x is even, i.e. x = 2x0.
Hence (5x0 ; 22y0 )(5x0 + 22y0 ) = 9, from where x0 = y0 = 1, i.e. x = 2 and y = 4.
Let t > 1. Then 16y0 1 (mod 25) and y0 is divisible by 5 (5 is the index of 3 modulo 25).
Hence y0 = 5y1 and 1 + 5x = 220y1 + 2:5t. On the other hand 210 1 (mod 11) and 5x 2:5t
(mod 11). Obviously x > t. Thus 5x;t 2 (mod 11), which is not true.
Therefore the given equation has a unique solution (x y z t) = (2 4 1 1).
Grade 12.
p 2Problem 1. For every real number x we denote by f (x) the maximal value of the function
t + 2t + 2 in the interval x ; 2 x].
a) Prove that f (x) is an even function and nd its minimal value
b) Prove that the function f (x) is not dierentiable for x = 0.
c) Prove that the sequence an = ff (n)g, n = 1 2 : : : is convergent and nd its limit.
(For every real number a we denote with fag the unique real number in the interval 0 1)
for which the number a ; fag is an integer.)
9
p
Solution. It is clear that the maximal value of the function t2 + 2t + 2 in the interval
x ; 2 x] is reached at the endpoints of this interval. Since the inequality (x ; 2)2 +2(x ; 2)+2
x2 + 2x + 2 is equivalent to x 0, then
8p 2
< x ; 2x + 2
> if x 0
f (x) = > p
: x2 + 2x + 2 if x 0:
a) Obviously f (x) = f (;x), i.e. f (x) is an even function and its minimal value is f (0) = 2.
b) Since p2 p p
lim f ( x) ; f (0) = lim x ; 2 x + 2 ; 2 = ; 2
x!0
x<0 x;0 x !
x<0
0 x 2
and p2 p p
lim f ( x ) ; f (0) = lim x + 2 x + 2 ; 2 = 2
x!0
x>0 x ; 0 x!0
x>0 x 2
then f 0(0) doesn't exist. np o p
c) It is clear that an = n2 + 2n + 2 . On the other hand n + 1 < n2 + 2n + 2 < n + 2,
p
i.e. an = n2 + 2n + 2 ; (n + 1).
Therefore
a = nlim p 1 = 0:
nlim
!1 n !1 n2 + 2n + 2 + n + 1
10
i.e. AP = AB 2AB:AC . Analogously DQ = 2DB:DC . These equalities show that the position
+ AC DC + DB
of the points P and Q doesn't depend on the circle k.
b) Let us denote AB = x, BC = y and CD = z . It follows from a) that
PQ = AD ; AP ; DQ = x + y + z ; 2xy(+x +2xy ) ; 2zy(+
y + z)
2z
= (x + y ) 1 ; y +2x2x + z 1 ; 2(yy++2zz)
= (yy+(x2x+)(yy ++ z2)z ) :
2
Therefore the minimal length of the segment PQ is b and this length is reached i AB =
2
a
CD = x = z = a ; b .
2
Problem 3. Find all prime numbers p and q , such that the number 2p + 2q is divisible by
p:q.
Solution.
Lemma. If k > 1, then k doesn't divide 2k;1 + 1.
Proof. Assume that k divides 2k;1 + 1. Obviously k is odd. Let k = p1 1 p2 2 : : :pr r , where
p1 p2 : : : pr are odd prime numbers (pi 6= pj if i 6= j ), 1 2 : : : r are positive integers and
r 1. Let pi ; 1 = 2mi :ti, where ti are odd integers ( i = 1 2 : : : r). Let m1 be the smallest
number in the sequence m1 m2 : : : mr . It follows from pi 1 (mod(pi ; 1)) that pi 1
(mod 2mi ) and pi i 1 (mod m1 :u
2mi ), i = 1 2 : : : r. Hence k ; 1 = 2m1 :u (u is an integer). If
2 ;1 (mod k), then 2
k ;1 2 ;1 (mod k) and 2(p1;1):u ;1 (mod p1), because t1 is
odd. But 2 1 1 (mod p1 ) | a contradiction.
p ;1
11
WINTER MATHEMATICAL COMPETITION
1996
Grade 8
Problem 1. For which integer values of the parameter a the equation j2x + 1j + jx ; 2j = a
has an integer solutions?
Solution. We shall consider the following cases:
I. x > 2. Then 2x + 1 > 0, x ; 2 > 0 and the equation is equivalent to 2x + 1 + x ; 2 = a.
Thus, x = 1 +3 a , which is a solution when 1 +3 a > 2, i.e. when a > 5.
II. ; 12 x 2. Then 2x+1 0, x;2 0 and the equation has the form 2x+1;(x;2) = a.
Thus, x = a ; 3, which is a solution when ; 1 a ; 3 2, i.e. 5 a 5. In this interval the
2 2
integers are a = 3 4 5 and respectively we get x = 0 1 2, which are integer solutions.
III. x < ; 21 . Then 2x+1 < 0, x;2 < 0 and the equation is equivalent to ;(2x+1);(x;2) =
a or ;3x = a ; 1. Thus x = 1 ;3 a , which is a solution, when 1 ;3 a < ; 12 , i.e. when a > 25 .
According to the considered cases the given equation has solution when a > 25 . When a = 3
the only integer solution is x = 0. When a = 4 there are two integer solutions x = 1 and x = ;1.
When a = 5, x = 2 is the only integer solution. When a > 5 the equation has two solutions
x1 = 1 +3 a and x2 = 1 ;3 a . For these solutions we have: if a = 3k ; 1, then x1 is integer only
if a = 3k + 1, then x2 is integer only if a = 3k, there is no integer solution.
As a result, we get integer solutions when a = 3 or a = 3k 1, where k is a positive integer,
greater than 1.
Problem 2. The bisector AD (D 2 BC ) of the acute isosceles triangle ABC divides it into
two isosceles triangles. Let O and I be the incenter and the circumcenter of 4ABC , respectively.
AO meets BC in point E , while F is the intersection point of the lines BI and DO. Prove that:
a) the quadrilaterals ABEF and ADCF are rhombi with equal side lengths
b) If H is the altitude center of 4ABE , then the points A, D, E , F , H are concyclic.
Solution. Firstly, let us justify the position of AD. If it is a bisector of the angle between the
two equal sides, then it is perpendicular to BC , 4ADB and 4ADC are isosceles. Consequently
AD = BD = CD and 6 BAD = 6 CAD = 45 , i.e. 6 BAC = 90, which contradicts to the
condition that 4ABC is acute. It follows that AD is bisector of the angle belonging to the base
AB, while 4ABD and 4ACD are isosceles (Figure 1).
a) If 6 BAD = 6 CAD = , then 6 ABC = 2. Since 6 ADB > 6 CAD = 6 BAD , the only
possibility is 6 ADB = 6 ABC = 2 and AD = AB . But 6 ACD = 6 ADB ; 6 CAD = 2 ; =
1
, i.e. 6 ACD = 6 CAD and CD = AD. From the equality 2 +2 + = 180 we dene = 36.
Thus, 6 ACB = 36, 6 BAC = 6 ABC = 72. The points O and I lie on the bisector of 6 ACB ,
which is a segment bisector of AB . We have OA = OC and 6 OAC = 6 OCA = 18 = 6 OAD,
while 6 OAB = 36 + 18 = 54, 6 AEB = 180 ; 72 ; 54 = 54 = 6 BAE . It follows that
BE = AB and the bisector BI of 6 ABC is the segment bisector of AE . In the triangle ACD,
which is isosceles, we have that AO and CO are bisectors and it follows that DO is the bisector
of 6 ADC and the segment bisector of AC . Then F is the intersection point of the segment
bisectors of AE and AC , i.e. F is the circumcenter of 4ACE , the radius of the circumcircle is
AF = EF = CF . We have 6 AFE = 26 ACE = 72 , while 6 AFB = 6 BFE = 36 = 6 ABF =
6 EBF and consequently AF = AB = BE = EF , i.e. ABEF is a rhombus. In the same way
CF = AF = AB = AD = CD and ADCF is a rhombus, which side is AD = AB .
b) It is clear that 6 AEF = 6 AEB = 54 and Figure 1.
1
6 ADF = 6 ADC = frac12(180 ; 72) = 54,
2
the segment AF is seen from the points D and E
under 54. We draw a line through A, perpendicular
to BC , which intersects BI in the point H | the
altitude center of 4ABE , because BI ?AE . But in
4ABD the altitude AH is a bisector of 6 BAD =
36. Thus, 6 BAH = 18, 6 AHF = 18 + 36 =
54 = 6 AEF = 6 AOF . Note that the points D, E ,
H are in one and the same semiplane with respect
to AF . Consequently, D, E , H together with A and
F are concyclic.
Problem 3. Every day a student preparing
himself for the Winter competition in mathemat-
ics has been solving problems during a period of 5
weeks. He has been solving at least one problem daily but no more than 10 problems weekly.
a) Prove that during some consecutive days the student has solved 19 problems exactly.
b) If 1 n 34 is a natural number, prove that during some consecutive days the student
has solved n problems exactly.
Remark. Every week begins on Monday and ends on Sunday.
Solution. Since a) is a particular case of b), we shall solve b) only.
According to the condition the student has been solving problems during 5 7 = 35 days and
has solved at most 5 10 = 50 problems. Let xi be the number of the problems, solved during
the i-th day (i = 1 2 : : : 35).
Let 1 n 34 be a xed natural number. We want to prove that there exist such k < l,
that xk+1 + + xl = n. Denote Xi = x1 + + xi . Obviously,
1 X1 < X2 < : : : < X35 50
and the problem is to prove the existence of such k < l, that Xl ; Xk = n.
Case 1. 1 n 19. We consider the numbers
X1 < X2 < : : : < X35 X1 + n < X2 + n < : : : < X35 + n (1)
which are integers and their number is 70. Obviously they are in the interval 1 50 + n], in
which there are 50 + n 50 + 19 < 70 integers. Consequently, among the numbers (1) there
2
are at least two equal. The rst 35 of them as well as the next 35 are dierent from each other.
Therefore, there exist such k and l, for which Xl = Xk + n, i.e. for which Xl ; Xk = n.
Case 2. 20 n 34. Firstly, we shall prove the following
Lemma. If the integers z1 z2 : : : zm belong to the interval 1 2n] and if m > n, then among
the numbers z1 z2 : : : zm there are two, the dierence of which is equal to n exactly.
Proof. With the numbers from the interval 1 2n] we construct the following pairs:
(1 n + 1) (2 n + 2) : : : (n 2n):
The number of these pairs is n and the dierence of the numbers in each pair is equal to
n. Since m > n, at least two of the numbers z1 z2 : : : zm belong to one and the same pair.
Therefore their dierence is equal to n.
Let us nish now the solution of the problem.
If n 25, then 2n 50, and thus all the numbers X1 X2 : : : X35 are in the interval
1 2n]. On the other hand n < 35 and according to the lemma there are two numbers among
X1 X2 : : : X35 which dierence is n.
If 20 n 24, we represent the interval 1 50] as an union of the intervals 1 2n] 2n +1 50].
In the second one there are 50 ; (2n + 1) + 1 = 50 ; 2n integers. Then, the number of the
integers among X1 X2 : : : X35, which belong to the interval 1 2n], is at least 35 ; (50 ; 2n) =
2n ; 15 40 ; 15 = 25 > n. Consequently, we can apply the lemma again.
Remark. The case 1 can be solved by the lemma proved above.
Grade 9
Problem 1. Let f (x) = x3 ; (p + 5)x2 ; 2(p ; 3)(p ; 1)x + 4p2 ; 24p + 36, where p is a real
parameter.
a) Prove that f (3 ; p) = 0.
b) Find all values of p, for which two of the roots of the equation
p f (x) = 0 are lengths of the
cathetuses of a rectangle triangle which hypotenuse is equal to 4 2.
Solution. a) We have f (x) = (x + p ; 3)(x2 ; 2(p p + 1)x + 4(p ; 3)).
b) The roots of f (x) = 0 are x12 = p + 1 p2 ; 2p + 13 and x = 3 ; p. If p > 3,
p 2 3
then x1 > 0, x2 > 0 and x3 < 0. The equation x21 + x22 = 32 = 4 2 gives p = 1, which is
impossible. If p = 3, then two of the roots are equal to 0 and this casepgives no solution. If p < 3,
then x1 > 0, x3 > 0 and x2 < 0. Therefore, 32 = x21 + x23 = (p + 1 + p2 ; 2p + 13)2 + (3 ; p)2.
This equation is equivalent to
q
(p + 1)(3(p ; 3) + 2 p2 ; 2p + 13) = 0:
p
If p 6= p;1, we get 2 p2p; 2p + 13 = 3(3 ;pp) > 0, from where 5p2 ; 46p + 29 = p 0 and
23 8 6 23 + 8 6 23 ; 8 6 23 ; 8 6.
p = 5 . Since > 3 and < 3, we nd p 1 = ;1 and p2 =
5 5 5
Problem 2. The incenter of the quadrilateral ABCD is O. The lines lA ? OA, lB ? OB ,
lC ? OC and lD ? OD are drawn through the points A, B, C and D respectively. The lines lA
and lB meet each other in the point K , lB and lC | in L, lC and lD | in M , lD and lA | in
N.
a) Prove that the lines KM and LN meet each other in the point O.
3
b) If the lengths of the segments OK , OL and OM are p, q and r respectively, nd the
length of the segment ON .
Solution. a) We shall prove that the points N , O and L are colinear. Denote 6 ABC = 6 B ,
6 BCD = 6 C , 6 CDA = 6 D and 6 DAC = 6 A. Since O is incenter, the segments OA,OB ,OC
and OD are the bisectors of the corresponding angles of the quadrilateral ABCD. Note that
each of the quadrilaterals AKBO, BLCO, CMDO and DNAO is inscribed.
Figure 2. Consequently: 6 NOK + 6 KOL = ;
6 ONA ; 6 OKA + ; 6 OKB ; 6 OLB =
; 6 ADO ; 6 6 ABO6 + ;6 6 BAO ; 6 BCO =
2 ; ( 2 + 2B + 2A + 2C ) = 2 ; = .
6 D
4( R1 + R2 )
and (2), we get
r = 4R(abc
2
a + b)2 : (3)
5
Since (a +abb)2 14 and c2 4R2, it follows from (3) that r R4 . The equation is reached
when a = b and c = 2R. Now, it follows that 6 A = 6 B = 45 and 6 C = 90.
Problem 3. A positive integer n and a Figure 4.
real number ' are given in a way that cos ' =
1
n . Find all positive integers k, for which the
number cos k' is an integer.
Solution. Case 1. n = 1. Then cos ' = 1
and ' = 2m (m = 0 1 : : :). For all k 2 N
the number cos k' is an integer.
Case 2. n = 2. Then cos ' = 21 and
' = 3 + 2m (m = 0 1 : : :). It is clear
that for all k 2 N, which are divisible by 3,
the number cos k' is an integer.
Case 3. n 3. We shall prove that for
all k 2 N the number cos k' is not an integer. Let n be odd. We have cos ' = n1 , cos 2' =
2 cos2 ' ; 1 = 2 ; n and (2 ; n2 n) = 1. We shall prove by induction that cos k' = a , where
2
n2 nk
(a n) = 1. Assume that the assertion is true for all integers from 1 to k. We shall check it for
k + 1. From
cos(k + 1)' + cos(k ; 1)' = 2 cos k' cos '
it follows that cos(k + 1)' = 2 a ; b = 2a ; bn , where cos k' = a and cos(k ; 1)' =
2
n nk nk;1 nk+1 nk
b . We have (a n) = 1 and (b n) = 1 according to the inductive assumption. It is clear that
nk;1
(2a ; bn2 n) = 1 and this ends the proof. The case when n is even is analogous. Now cos k'
is expressed by a fraction which denominator is equal to 2pk with n = 2p, while the nominator
has no common divisor with p.
Answer: if n = 1 8k 2 N
if n = 2 k = 3q , where q 2 N
if n 3 there is no solution.
Grade 11
Problem 1. Find the values of the real parameter a, for which the function
f (x) = x2 ; 2x ; jx ; 1 ; aj ; jx ; 2j + 4
has nonnegative values for all real x.
Solution. Firstly let 1 + a 2, i.e. a 1. Then
8
>
< x2 + 1 ; a x 1+a
f (x) = > x2 ; 2x + 3 + a 1 + a x 2
: x2 ; 4x + 7 + a x 2:
6
If 1 + a 2, i.e. a 1, we nd that
8 2
< x + 1 ; a
> x2
f (x) = > x2 ; 2x + 5 ; a 2 x 1 + a
: x2 ; 4x + 7 + a x 1 + a:
PB PO
= sin(90
sin(2 ; x) ; )
QA QO
= sin(90
sin(180 ; x) ;
)
QC QO :
= sin(90
sin(2
; 180 + x) ;
)
Then
PB sin 2
+ QC sin 2 = sin(2 ; x) sin 2
; sin(2
+ x) sin 2 =
PA QA sin x sin x
1 ((sin 2 cos x ; sin x cos 2 ) sin2
; (sin 2
cos x + sin x cos 2
) sin 2 )
sin x
= ; sin 2
cos 2 ; sin 2 cos 2
= ; sin 2(
+ ) = sin 2:
Conversely, let the given equality be satised. Denote by Q0 0the intersection point of OP
and AC . It follows from the above that sin 2 = PA sin 2
+ Q0 A sin 2 . Then Q0A = QC
PB Q C Q 0C
QA .
Since the points Q0 and Q lie on the segment AC , then Q0 = Q.
Problem 3. Find all functions f (x) with integer values and dened in the set of the integers,
such that
3f (x) ; 2f (f (x)) = x
7
for all integers x.
Solution. The function f (x) = x satises the condition of the problem.
Let f (x) be a function which satises the condition. Let g (x) = f (x) ; x. The condition can
be written in the form
2f (f (x)) ; 2f (x) = f (x) ; x
which is equivalent to
g(x) = 2g (f (x)):
From here we obtain
g(x) = 2g (f (x)) = 22 g (f (f (x))) = 23g (f (f (f (x)))) = 24g (f (f (f (f (x))))) = : : ::
Since the numbers g (f (f : : : f (x)) : : :) are integer, then g (x) is divisible by 2n for all integers
x and all natural numbers n. This is possible only if g (x) = 0. Thus, f (x) = x is the only
solution of the problem.
8
Winter mathematics
competition|Burgas, 1997
Problem 8.1. Let F be the set of points with coordinates (x y)
such that jjxj ; jyjj + jxj + jyj = 2.
(a) Draw F .
(b) Find the number of points in F such that 2y = j2x ; 1j ; 3.
Solution: (a) If jxj jyj, then jjxj ; jyjj + jxj + jyj = jxj ; jyj +
jxj+jyj = 2jxj = 2, thus jxj = 1 and therefore 1 jyj, so ;1 y 1.
We conclude that the segments ;1 y 1 on the lines x = 1 and
x = ;1 belong to F .
If jxj jyj, then jjxj ; jyjj + jxj + jyj = ;jxj + jyj + jxj + jyj =
2jyj = 2, thus jyj = 1 and therefore 1 jxj, so ;1 x 1. We
conclude that the segments ;1 x 1 on the lines y = 1 and
y = ;1 also belong to F .
Thus we have determined that F consists of the sides of a square
with vertices A(;1 ;1) B (1 ;1) C (1 1) D(;1 1).
1
(b) We nd the number of solutions of 2y = j2x ; 1j; 3 on each
of the segments AB BC CD DA.
The segment CD consists of all points (x y) such that ;1 x
1, y = 1. The equation 2 = j2x ; 1j ; 3 has no solution x when
;1 x 1. Therefore 2y = j2x ; 1j ; 3 has no solution on CD.
The segment AB consists of all points (x y) such that ;1 x
1, y = ;1. The equation ;2 = j2x ; 1j ; 3 has two solutions: x = 0
and x = 1. Therefore 2y = j2x ; 1j ; 3 has two solutions on AB .
As above we get that 2y = j2x ; 1j ; 3 has a unique solution
on AD: (x y) = (;1 0) and a unique solution on BC : (x y) =
(1 ;1). Note that the last one has already been obtained as a point
on AB . Thus there are three solutions of 2y = j2x ; 1j ; 3 in F :
(x y) = (;1 0) (0 ;1) (1 ;1).
2
Solution: Denote by M , P and N the midpoints of AB and CH
and the intersecting point of the internal bisectors of 6 CAH and
6 CBH (g. 1).
Problem 8.3. The n points A0 A1 : : : An;1 lie a circle in this
order and divide it into equal arcs. Find an ordering B0 B1 : : : Bn;1
3
of the same points such that the length of B0B1 : : :Bn;1 is maximal.
Thus
(1) pqn = (;1)n+1 :
It follows from (1) for n = 1 and n = 2 that pq = 1 and pq2 = ;1
and so p = ;1, q = ;1. Direct verication shows that p = ;1 and
q = ;1 satisfy (1) for any n. Also 6= .
(b) Since and are the roots of x2 ; x ; 1 = 0, we know that
= + 1 and 2 = + 1. Therefore
2
8
Problem 10.1. Find all real numbers x such that tan 12 ; x ,
tan 12 and tan 12 + x form (in some order) a geometric progres-
sion.
Solution: Denote a = tan 12 and y = tan x. There are three cases
to consider:
1. tan 12 ; x tan 12 + x = tan2 12 . Now 1a+;ayy 1a;+ayy =
a2. Therefore a2 ; y2 = a2(1 ; a2y2), and so (a4 ; 1)y2 = 0.
Since a 6= 1 we get that y = 0, so tan x = 0. Obviously
all numbers x of the kind x = k, k 2 Zare solutions to the
problem.
2. tan 12 tan 12 + x = tan 12 ; x . We obtain
2
!2
a + y a ; y
a 1 ; ay = 1 + ay =) (a2 + 1)yay2 + (a2 ; 1)y + 3a] = 0:
The case of y = 0 is settled in 1. Let y1 and y2 be the roots of
the equation ay2 + (a2 ; 1)py + 3a = 0. Since a = p tan 15 deg =
tan(45 deg ;30 deg) = 2 ; 3 we get y1 = y2 = 3, so tan x =
p
3. Obviously all x of the kind x = 3 + k, k 2 Zare solutions
of the problem.
3. tan 12 tan 12 ; x = tan 12 + x . The substitution z =
2
Problem 10.2. Two points C and M are given in the plane. Let
H be the orthocentre of 4ABC such that M is a midpoint of AB .
(a) Prove that CH CD = jAM 2 ; CM 2j where D 2 AB and
CD ? AB:
(b) Find the locus of points H when AB is of given length c.
10
so
(2) AM 2 ; CM 2 = CD(DH ; CD):
In case 1 we have CD = CH + DH and from (2) we get
AM 2 ; CM 2 = CD(DH ; CH ; DH ) = ;CH CD:
In case 2 we have CD = DH ; CH and from (2) we get
AM 2 ; CM 2 = CD(DH ; DH + CH ) = CH CD:
In case 3 it follows from (1) that CD DH = (DM
AM )(DM
AM ) = DM 2 ; AM 2 = CM 2 ; CD2 ; AM 2 and therefore CM 2 ;
AM 2 = CD(DH + CD). Now DH = CH ; CD and so CM 2 ;
AM 2 = CD(CH ; CD + CD) = CH CD. Summarising all three
cases considered (and the case of a right triangle 4ABC ) we get
CH CD = jAM 2 ; CM 2j:
11
(b) It follows from the conditions of the problem that A and B
are diametrically opposite in the circle (k) of centre M and radius
c . When the diameter AB varies in (k) we obtain all triangles ABC
2
with the xed vertex C . There are three cases:
12
Consider now case 2. Let AB be a diameter in (k) perpendicular
to CM . The orthocentre H of 4ABC lies on line CM . Moreover
H is an external point for (k) because 4ABH is acute. The last
follows from 6 AHB = 180 deg ;6 ACB and 6 ACB > 90 deg.
Let A1B1 be a arbitrary diameter of (k) such that C 2= A1B1:
Denote by H1 the orthocentre of A1B1C and by D1 the foot of the
altitude from the vertex C . Since 4A1B1C is obtuse, we get that
H1 6= C . Consider 4MD1C and 4CHH1 and use a). We are in
case 2 of (a). For 4ABC we obtain that CH CM = c4 ; CM 2 and
2
CM
so 4MD1C 4CHH1. Since 4MD1C is a right triangle, we
1
13
and it follows from the above that it lies on l. Since H1 is the
intersecting point of these two lines it is obvious that H1 is the
orthocentre of 4A1B1C .
Therefore the locus is a line l perpendicular to CM through a
point H on the ray opposite to CM~ and of distance 4jCM c2 ;jCM j
j
from point C .
Consider case 3. Let AB be a diameter of (k) perpendicular to
CM . The orthocentre H of 4ABC lies on CM but now H is an
internal point for (k). Consider an arbitrary diameter A1B1 of (k)
such that H 2 A1B1. Let H1 be the orthocentre of 4A1B1C and
CD1{ the altitude. Consider 4MD1C and 4CHH1. It is essential
to show that 4CHH1 is uniquely determined (for 4MD1C it is
obvious). It suces to prove that H1 6= C . If the contrary is true,
then 4A1B1C is a right triangle with its right angle at C and since
A1B1 is a diameter, C must lie on (k), a contradiction. Since 4ABC
is acute, we apply case 1 of (a), so CH CM = CM 2 ; c4 . There
2
14
The further considerations follow those from case 2. We conclude
that the locus is a line perpendicular to CM and passing through
H 2 CM~ of distance jCM j ; 4jCM c2 from C .
j
In particular it follows that PQ?CM , a well-known property.
Problem 10.3. How many natural numbers a1a2 : : : a2n exist such
that:
(a) none of the digits ai is zero
(b) the sum a1a2 + a3a4 + : : : + a2n;1a2n is an even number?
15
Problem 11.1. The sequence fxng1n=1 is dened as:
x1 = 3 xn+1 = x2n ; 3xn + 4 n = 1 2 3 : : :
(a) Prove that fxng1
n=1 is monotone increasing and unbounded.
16
Since 0 x 1 ; 2 n1 it follows that nlim 1
!1 xn+1 ; 2
= 0, so
n+1
nlim y = 1.
!1 n
18
Solution: Let m and n satisfy the conditions of the problem. Since
n is an odd integer number, we get (m n) = 1 and n > 2. When
n = 3, all m 4 such that m 1(modn 3) are solutions, because
if m ;1(mod 3), then 1 + m +n m23 1 ; 1 + 1 1(mod 3).
3n
Let nown n > 3. nIt follows that m3 6 1(mod n), because otherwise
1 + m3 + m23 n+13(mod n), i. e., n=3. On the other hand 1 +
m3n + m23n = mm3n ;;11 and therefore m3n+1 1(mod n). Let k
3
19
Winter mathematics
competition|Pleven, 6{8
February 1998
Dedicated to the One Hundredth Anniversary of the UBM
Solution:
a.) It is obvious that a 6= 0, b 6= 0, c 6= 0. The rst equality
gives b2 = ac, whence by multiplying both sides by b we get
b3 = abc. Similarly a3 = abc and c3 = abc. Hence a3 = b3 = c3
and therefore a = b = c.
b.) By multiplying both the numerator and the denominator of
the second fraction by 3 and using the result of a.) we obtain
x = 3y. Thus ;4x2 + 36y ; 8 = ;9(4y2 ; 4y + 1) + 1 =
;9(2y ; 1)2 + 1, and its maximum value is 1 when 2y ; 1 = 0.
Therefore y = 21 and x = 23 , i. e., x + y = 2.
2
Solution:
a.) The segments EM and FM are medians to the hypotenuses
of 4BCE and 4BCF and therefore EM = FM = 21 BC .
Similarly, for 4AHE and 4AHF we get EK = FK = 12 AH .
Since 6 BAC = 45 deg, we nd that 4AEB and 4CEH are
isosceles. Hence AE = BE and EC = EH , i. e., 4AHE =
4BCE . Therefore EK = EM . Thus MEKF is a rhombus.
Furthermore,
6 MEK = 6 MEB + 6 HEK = 6 CBE + 6 HEK
b.) It follows from a.) that the intersecting point S of the diagonals
of the quadrangle MEKF is the midpoint of both diagonals.
Since 4AEB is isosceles, E lies on the axis of symmetry of the
segment AB and therefore EO?AB , i. e., EOkHF . Similarly
FOkEH . Thus the quadrangle EOFH is a parallelogram.
From the above we conclude that S is the midpoint of OH .
3
c.) a.) implies that in the acute triangle 4ABC with orthocentre
H and 6 BAC = 45 deg it is true that AH = BC . 4AFE
is of the same type and therefore EF = AO = 1. (It follows
from b.) that O is orthocentre of this triangle.)
Problem 8.3. Let 1998 points be chosen on the plane so that out of
any 17 it is possible to choose 11 that lie inside a circle of diameter 1.
Find the smallest number of circles of diameter 2 sucient to cover
all 1998 points.
(We say that a circle covers a certain number of points if all
points lie inside the circle or on its outline.)
4
Now choose a circle of radius 1 centred in one of the points. If
the remaining points lie inside the circle, the required number is 1
and thus no greater than 7. If this is not the case, take another
point outside the rst circle. If all points lie in the two circles, then
the required number is 2 and thus no greater than 7. Continuing in
this way we either obtain no more than 7 circles covering all points
or have 7 circles and a point that lies outside all circles. Consider
this point and the centres of the chosen circles. There exist 2 points
among these 8 such that the distance between them is no greater
than 1. But this is impossible because of the way we chose our
points.
Together the two parts of the proof demonstrate that the required
number is 7.
5
Subtracting the two equalities
m2 ; am + b = 7
p2 ; ap + b = ;7
we nd
14 = m2 ; p2 ; a(m ; p) = (m ; p)(m + p ; a) = (m ; p)(p ; n):
Thus the numbers m ; p and p ; n are either both positive or both
negative and since m > n, they are positive. Moreover they are
integer and therefore are equal to 1 and 14 or to 2 and 7. But since
m n p 2 1 9] it follows that neither m ; p nor p ; n is 14. There
are two cases to consider: m ; p = 2, p ; n = 7 and m ; p = 7,
p ; n = 2, i. e., either m = p + 2, n = p ; 7 or m = p + 7, n = p ; 2.
It is obvious that in both cases at least one of m n p lies outside the
interval 1 9].
Case 2. f (m) = f (n) = ;7, f (p) = 7. As in Case 1 we get
a = m + n, b = mn ; 7 and (m ; p)(p ; n) = ;14. Using similar
arguments we obtain that either m ; p = 2, p ; n = ;7 or m ;
p = ;2, p ; n = 7. (Without loss of generality we suppose that
jm ; pj < jp ; nj.) Therefore the two options are m = p + 2,
n = p + 7 and m = p ; 2, n = p ; 7. Simple calculations show that
all triples (m n p) satisfying the conditions are (3 8 1), (4 9 2),
(6 1 8) and (7 2 9). So the functions are f (x) = x2 ; 11x + 17,
f (x) = x2 ; 13x + 29, f (x) = x2 ; 7x ; 1 and f (x) = x2 ; 9x + 7.
8
Problem 10.1. Find all values of the real parameter a for which
the equation x3 ; 3x2 +(a2 +2)x; a2 = 0 hasthree distinct roots
x1,
x2 and x3 such that sin 23 x1 , sin 23 x2 and sin 23 x3 form
(in some order) an aritmetic progression.
10
and after dividing by cos 2 cos (cos 2 6= 0 (why?), and if cos = 0,
i. e., = 90 , we have one of the two cases already considered)
we obtain
tan 2 (1 + 2 tan ) = 2 + tan
It follows in particular that if is xed, then is uniquely deter-
mined.
Suppose that tan 2 > 2. Thus 2 > 45 and therefore > 90.
If tan < 0, we get > 90 , which is impossible, since + < 180.
If tan > 0, we get 2 + tan > (1 + 2 tan ) 2 i. e., tan < 0, which
is a contradiction.
Therefore tan 2 2 and B lies on CED where 6 GCF = 6 FCD
and tan 6 GCF = 2 (g. 1).
Fix the point B such that < 90. Let T be the midpoint of the
arc CB and let A1 be the intersecting point of TF and the circle.
We shall show that A1
A. We obtain 6 OA1F = 2 ; (90 ; ) =
+ 2 ; 90 and 6 A1FO = 45 ; 2 + ; 45 = ; 2 . It follows from
11
sin ; 2
the Sine Theorem for 4A1FO that = 2, which is
sin
2+ ; 90
i i i i
also a2k ;1a2k ) are uniquely determined by a2k1;1 + a2k1 (there are
i i
two possibilities for a2k1;1 a2k1 ). There are two possibilities for any
of the remaining n ; s pairs (for which a2t;1 + a2t = 0). Therefore
there are
! ! ! ! !
;1 n ;2 n ;k n n n :
2 +22 1 +22 2 + +22
n n n n
k + : : : +2 2 n;1 +2 n
12
sequences with the required property. By adding and subtracting 2n
to and from the above expression we get:
! ! ! ! !!
2: 2 0 +2 1 +2 2 + +2 n ; 1 + nn ;2n =23n ;2n
n n n;1 n n ;2 n n
Solution: a.) By grouping the rst and third radicals and the
second and fourth radicals and rationalising we get that when x > 4,
f (x) = px+p1 x;1 ; px;4+1 px;3 . Therefore xlim
!1
f (x) = 0.
b.) When x > 4,
f 0(x) = 2p1 x ; p 1 ; p 1 + p 1
2 x;1 2 x;3 2 x;4
= p p 1
p p +
2 x;3 x;4( x;3 + x;4)
; 2pxpx;1(1px+px;1) > 0:
13
Therefore f (x) is an increasing function if x 4.
c.) It follows from a.) and b.) that f (x) < 0 when x 4, i. e.,
the equation couldq have a q solution only if a < 0. Let a < 0. The
function g(x) = a x = a 1 ; x3 is decreasing and continuous and
x ;3
xlim
!1
g(x) = a < 0. Since f (x) is an increasing and continuous func-
tion and xlim
!1
f (x) = 0, in accordance with the Bolzano{Weierstra
Theorem the equation f (x) = g(x) has a solution
p (and it is a unique
one) exactly when f (4) g(4), i. e., if 2(1 ; 3) a < 0.
14
Note: The above solution shows that if O is the intersecting
point of the axes of symmetry of AC and BD, then the assertion of
the problem and its opposite are true for a quadrangle that is not
inscribed in a circle. This could be demonstrated by using complex
numbers or trigonometry.
15
Solution: Let m be an integer number such that m = b0 + b1 5 +
+ bn 5n . Denote f (m) = 2b0 + 2b1 10 + + 2bn 10n . It is clear
that ff (m) j m 2 Ng is the set of integer numbers with only even
digits in their decimal representation. Since f (m1) < f (m2) ()
m1 < m2, it follows that am = f (m) for any m. Therefore it suces
to nd all m such that
12(b0 + b1 5 + + bn 5n ) = 2b0 + 2b1 10 + + 2bn 10n
i. e.,
(1) 6(b0 + b1 5 + + bn 5n ) = b0 + b1 10 + + bn 10n :
Since b0+b15+ +bn5n 5n+1 ;1 and b0+b110+ +bn 10n 10n ,
it follows from (1) that 6(5n+1 ; 1) 10n , i. e., 6 5n+1 > 10n . Thus
2n < 30 and therefore n 4. If n = 4, we get from (1) that
b0 + 4b1 + 10b2 = 50b3 + 1250b4 1250, which is imposible. In the
same way it is easy to show that n 3, i. e., n = 3. In this case
b0 +4b1 +10b2 = 50b3. Obviously b3 = 1 and b0 = b1, because b0 ;b1 is
divisible by 5. As a result we have the equation b0 +2b2 = 10, and its
solutions are b0 = 2 b2 = 4 and b0 = 4 b2 = 3. Therefore all integer
numbers m with the required property are m = 2+2 5+4 52 +53 =
237 and m = 4 + 4 5 + 3 52 + 53 = 224.
16
Winter mathematics
competition|Varna, 1999
Problem 8.1. Find all natural numbers x and y such that:
a) x1 ; y1 = 13
b) x1 + y1 = 13 + xy
1.
1
problem has four solutions:
x = 4 y = 9 x = 5 y = 6 x = 6 y = 5 x = 9 y = 4:
2
Let M4 and G4 be the projections of M and G on NN3 . It easily
follows now that
NN3 ; GG3 = GN = GG2 :
NN3 ; MM3 NM MM2
Further, using that NN3 = NN1 and MM3 = MM2, we obtain
GG3 MM2 = GG1 MM2 + GG2 MM2 and therefore GG3 =
GG1 + GG2.
Conversely, let the altitude through C be the sum of the remain-
ing two. Now GG3 = GG1 + GG2. If G? = GG3 \ MN , then it
follows straightforwardly that the sum of the distances from G? to
AC and BC equals to G?G3. It is easy to check now that G? = G.
1 2
4
Solution: a) It follows from the equation that x22 = 3px2 + p and
so 3px1 + x22 ; p = 3p(x1 + x2) = 9p2 > 0. The inequality is strict
because otherwise x1 = x2 = 0.
b) As in a), we obtain 3px1 + x22 + 3p = 3px2 + x21 + 3p =
9p2 + 4p > 0 (the last inequality follows from the conditions of the
problem x1 and x2 to be distinct and real, giving p 6= 0). Therefore
A = 9p2p+ 4p + 9p p+2 4p 2
2 2
5
b) Let L = CM \ k1. It follows from
a) that ML MC = MP 2 = MD MR.
We conclude that L lies on the circum-
cicrle of 4DRC and therefore RL ?
CM . Further HL ? CM , since HC is
a diameter of k1. Hence RH ? CM .
a2 sin x
2
7
" p p#
The solution of this inequality is a 2 ; 1 ; 5 ;1 + 5 . So we
" p# 2 2
obtain a 2 1 ;1 + 5 .
2
3 3
beta
2 sin 3 cos 3 = 1 () cos = 1 :
() coscos sin 3 cos 2 cos 3
8
It follows from Ceva's Theorem for point Q that:
(2)
AM BR CS = 1 () c SBSC = 1 () c BS a sin 3 = 1
MB RC SA b SABS b BS c sin 23
c a sin a sin a
() b c sin 2 = 1 () 2b sin cos
3 3
= 1 () b = 2 cos 3 :
3 3 3
Now (1) and (2) imply cos = b . From the Sine Law we obtain
cos a
sin = b , so cos = sin () sin 2 = sin 2 . If = , then
sin a cos sin
the triangle is isosceles and therefore P Q, implying that 6 ABP =
6 PBQ = 6 QBC = 0 , which is impossible. Thus + = 90 and
therefore 6 ACB = 90 .
b) It follows from 4BCS that cos = a . Combining the
3 BS
above with (2) gives 2b = BS . Note that 4ABC 4BHC , which
BP = BS = 2. Therefore BP = 2CH .
implies CH AC
Problem 10.3. Let A be a set of natural numbers with no zeroes in
their decimal representation. It is known that if a = a1a2 : : :ak 2 A,
then b = b1b2 : : :bk , where bj 1 j k is the remainder of 3aj
modulo 10, belongs to A and the sum of the digits of b equals the
sum of the digits of a.
a) Prove that the sum of the digits of a k-digit number in A equals
5k.
9
b) Find the smallest k-digit number which could be an element
of A.
11
an for any n. This inequality is equivalent to
(?) a(2n + 1) ;1:
q p
(n + 1)2 + 1 + n2 + 1
Since
2n + 1 2 + 1
nlim
!1
q p = lim q n q =1
(n + 1)2 + 1 + n2 + 1 n!1 (1 + n1 )2 + n12 + 1 + n12
it follows from (?) that a ;1.
Conversely, let a ;1. It follows from q 2n + 1
p <
(n + 1)2 + 1 + n2 + 1
2n + 1 = 1 that (?) holds true so the sequence fa g1 is increas-
n n=1
n+1+n
ing. The required values of a are a 2 ;1 +1).
12
We obtain B a B3
B2B3 = tg = r
2 p;a
and Bc B3 = tg 2 = p ;r c , so
B B
2 3
BaB3 = p ; c . Since B B = b ,
a c
BcB3 p ; a 2
it follows that BaB3 = 2 = p ; c
CBc and B B = p ; a = ABa ,
c 3
2 2 2
which gives AB3 = CB3. There-
fore B2O = B2B3 + B3O =
(p ; c)(p ; a) + R cos .
2r
We shall show that the above expression equals R + 2r and by analogy
A2O = C2O = R + 2r , which will complete the proof.
We obtain that (p ; c2)(rp ; a) +R cos = 2r +R () 2(pS; b) ;
S = R(1 ; cos ) () Sb = 2R sin2 () r =
2p 2p(p ; b) 2 p;b
2
4R sin2 () tg = 2 sin 2 () sin = 2 sin cos , which is
b 2 2 sin 2 2
a true equality.
Problem 11.3. Find the smallest natural number n such that the
sum of the squares of its divisors (including 1 and n) equals (n +3)2.
Solution: It is clear that n has at least three divisors and let 1 <
d1 < d2 < < dk < n be those dierent from 1 and n. The
13
conditions of the problem imply
(?) d21 + d22 + + d2k = 6n + 8:
Let n = p, where p is a prime number. It follows now from (?) that
p2 + p4 + + p2;2 = 6p + 8, so pn8 and therefore p = 2. The
above equality implies 1 + p2 + p4 + + p2;4 = 6p;2 + 2, which
is impossible.
Therefore k 6= 1 3 5, because otherwise the number of divisors
of n equals 3, 5, 7, i. e., n = p2 , n = p4 or n = p6, where p is a
prime number. Suppose that k 6. Since didk;i = n, it follows
from (?) that (dk;1 ; d1)2 + (dk;2 ; d2)2 + (dk;3 ; d3)2 8. The
last inequality is impossible, since the numbers dk;1 ; d1 dk;2 ; d2
and dk;3 ; d3 are distinct (if for example dk;1 ; d1 = dk;2 ; d2 = A,
then d1(A + d1) = d2(A + d2), so d1 = d2). We conclude now that
k = 2 or k = 4.
Assume k = 4. Then n has 6 divisors and thus n is of the form
n = p q2, where p and q are distinct prime numbers. (n is not of
the form n = p5). It follows from (?) that
(??) p2 + q2 + q4 + p2 q2 = 6pq2 + 8:
If q 5, then q4 + p2q2 2pq3 10pq2 > 6pq + 8 and therefore
q = 2 or q = 3. Direct verication shows that inequality (??) is
impossible. Thus k = 2 and hence n = pq, where p and q are
distinct prime numbers such that
p2 + q2 = 6pq + 8:
Since qnp2 ; 8, it is easy to see that if p 17 then p = 7, q = 41
and n = 287. Since 172 = 289 > 287, we conclude that the smallest
n with the required property is n = 287.
14